From ea70134ed3891934b262b3fd140354b1d9098c31 Mon Sep 17 00:00:00 2001 From: k8s-ci-robot Date: Thu, 17 Sep 2020 02:20:06 +0000 Subject: [PATCH] Deploy GitHub Pages --- .../20190815-zone-aware-routing/index.html | 61 ++++++----- search/search_index.json | 2 +- sitemap.xml | 100 +++++++++--------- sitemap.xml.gz | Bin 687 -> 687 bytes 4 files changed, 86 insertions(+), 77 deletions(-) diff --git a/enhancements/20190815-zone-aware-routing/index.html b/enhancements/20190815-zone-aware-routing/index.html index a96c567bc..87a6a8f3c 100644 --- a/enhancements/20190815-zone-aware-routing/index.html +++ b/enhancements/20190815-zone-aware-routing/index.html @@ -1186,19 +1186,19 @@

Summary

Teach ingress-nginx about availability zones where endpoints are running in. This way ingress-nginx pod will do its best to proxy to zone-local endpoint.

Motivation

-

When users run their services across multiple availability zones they usually pay for egress traffic between zones. Providers such as GCP, Amazon EC charges money for that. -ingress-nginx when picking an endpoint to route request to does not consider whether the endpoint is in different zone or the same one. That means it's at least equally likely -that it will pick an endpoint from another zone and proxy the request to it. In this situation response from the endpoint to ingress-nginx pod is considered as -inter zone traffic and costs money.

-

At the time of this writing GCP charges $0.01 per GB of inter zone egress traffic according to https://cloud.google.com/compute/network-pricing. -According to https://datapath.io/resources/blog/what-are-aws-data-transfer-costs-and-how-to-minimize-them/ Amazon also charges the same amount of money sa GCP for cross zone, egress traffic.

+

When users run their services across multiple availability zones they usually pay for egress traffic between zones. Providers such as GCP, and Amazon EC2 usually charge extra for this feature. +ingress-nginx when picking an endpoint to route request to does not consider whether the endpoint is in a different zone or the same one. That means it's at least equally likely +that it will pick an endpoint from another zone and proxy the request to it. In this situation response from the endpoint to the ingress-nginx pod is considered +inter-zone traffic and usually costs extra money.

+

At the time of this writing, GCP charges $0.01 per GB of inter-zone egress traffic according to https://cloud.google.com/compute/network-pricing. +According to https://datapath.io/resources/blog/what-are-aws-data-transfer-costs-and-how-to-minimize-them/ Amazon also charges the same amount of money as GCP for cross-zone, egress traffic.

This can be a lot of money depending on once's traffic. By teaching ingress-nginx about zones we can eliminate or at least decrease this cost.

-

Arguably inter-zone network latency should also be better than cross zone.

+

Arguably inter-zone network latency should also be better than cross-zone.

Goals

Non-Goals

Proposal

-

The idea here is to have controller part of ingress-nginx to (1) detect what zone its current pod is running in and (2) detect the zone for every endpoints it knows about. -After that it will post that data as part of endpoints to Lua land. Then Lua balancer when picking an endpoint will try to pick zone-local endpoint first and -if there is no zone-local endpoint then it will fallback to current behaviour.

-

This feature at least in the beginning should be optional since it is going to make it harder to reason about the load balancing and not everyone might want that.

+

The idea here is to have the controller part of ingress-nginx +(1) detect what zone its current pod is running in and +(2) detect the zone for every endpoint it knows about. +After that, it will post that data as part of endpoints to Lua land. +When picking an endpoint, the Lua balancer will try to pick zone-local endpoint first and +if there is no zone-local endpoint then it will fall back to current behaviour.

+

Initially, this feature should be optional since it is going to make it harder to reason about the load balancing and not everyone might want that.

How does controller know what zone it runs in? -We can have the pod spec do pass node name using downward API as an environment variable. -Then on start controller can get node details from the API based on node name. Once the node details is obtained -we can extract the zone from failure-domain.beta.kubernetes.io/zone annotation. Then we can pass that value to Lua land through Nginx configuration +We can have the pod spec pass the node name using downward API as an environment variable. +Upon startup, the controller can get node details from the API based on the node name. +Once the node details are obtained +we can extract the zone from the failure-domain.beta.kubernetes.io/zone annotation. +Then we can pass that value to Lua land through Nginx configuration when loading lua_ingress.lua module in init_by_lua phase.

How do we extract zones for endpoints? We can have the controller watch create and update events on nodes in the entire cluster and based on that keep the map of nodes to zones in the memory. And when we generate endpoints list, we can access node name using .subsets.addresses[i].nodeName and based on that fetch zone from the map in memory and store it as a field on the endpoint. -This solution assumes failure-domain.beta.kubernetes.io/zone annotation does not change until the end of node's life. Otherwise we have to +This solution assumes failure-domain.beta.kubernetes.io/zone annotation does not change until the end of the node's life. Otherwise, we have to watch update events as well on the nodes and that'll add even more overhead.

-

Alternatively, we can get the list of nodes only when there's no node in the memory for given node name. This is probably a better solution -because then we would avoid watching for API changes on node resources. We can eagrly fetch all the nodes and build node name to zone mapping on start. -And from thereon sync it during endpoints building in the main event loop iff there's no entry exist for the node of an endpoint. +

Alternatively, we can get the list of nodes only when there's no node in the memory for the given node name. This is probably a better solution +because then we would avoid watching for API changes on node resources. We can eagerly fetch all the nodes and build node name to zone mapping on start. +From there on, it will sync during endpoint building in the main event loop if there's no existing entry for the node of an endpoint. This means an extra API call in case cluster has expanded.

How do we make sure we do our best to choose zone-local endpoint? -This will be done on Lua side. For every backend we will initialize two balancer instances: (1) with all endpoints -(2) with all endpoints corresponding to current zone for the backend. Then given the request once we choose what backend -needs to serve the request, we will first try to use zonal balancer for that backend. If zonal balancer does not exist (i.e there's no zonal endpoint) -then we will use general balancer. In case of zonal outages we assume that readiness probe will fail and controller will -see no endpoints for the backend and therefore we will use general balancer.

+This will be done on the Lua side. For every backend, we will initialize two balancer instances: +(1) with all endpoints +(2) with all endpoints corresponding to the current zone for the backend. +Then given the request once we choose what backend +needs to serve the request, we will first try to use a zonal balancer for that backend. +If a zonal balancer does not exist (i.e. there's no zonal endpoint) +then we will use a general balancer. +In case of zonal outages, we assume that the readiness probe will fail and the controller will +see no endpoints for the backend and therefore we will use a general balancer.

We can enable the feature using a configmap setting. Doing it this way makes it easier to rollback in case of a problem.

Implementation History

Drawbacks [optional]

More load on the Kubernetes API server.

diff --git a/search/search_index.json b/search/search_index.json index 64f9740f1..9874463ef 100644 --- a/search/search_index.json +++ b/search/search_index.json @@ -1 +1 @@ -{"config":{"lang":["en"],"min_search_length":3,"prebuild_index":false,"separator":"[\\s\\-]+"},"docs":[{"location":"","text":"Welcome \u00b6 This is the documentation for the NGINX Ingress Controller. It is built around the Kubernetes Ingress resource , using a ConfigMap to store the NGINX configuration. Learn more about using Ingress on k8s.io . Getting Started \u00b6 See Deployment for a whirlwind tour that will get you started.","title":"Welcome"},{"location":"#welcome","text":"This is the documentation for the NGINX Ingress Controller. It is built around the Kubernetes Ingress resource , using a ConfigMap to store the NGINX configuration. Learn more about using Ingress on k8s.io .","title":"Welcome"},{"location":"#getting-started","text":"See Deployment for a whirlwind tour that will get you started.","title":"Getting Started"},{"location":"development/","text":"Developing for NGINX Ingress Controller \u00b6 This document explains how to get started with developing for NGINX Ingress controller. Prerequisites \u00b6 Install Go 1.14 or later. Note The project uses Go Modules Install Docker Important The majority of make tasks run as docker containers Quick Start \u00b6 Fork the repository Clone the repository to any location in your work station Add a GO111MODULE environment variable with export GO111MODULE=on Run go mod download to install dependencies Local build \u00b6 Start a local Kubernetes cluster using kind , build and deploy the ingress controller make dev-env Testing \u00b6 Run go unit tests make test Run unit-tests for lua code make lua-test Lua tests are located in the directory rootfs/etc/nginx/lua/test Important Test files must follow the naming convention _test.lua or it will be ignored Run e2e test suite make kind-e2e-test To limit the scope of the tests to execute, we can use the environment variable FOCUS FOCUS=\"no-auth-locations\" make kind-e2e-test Note The variable FOCUS defines Ginkgo Focused Specs Valid values are defined in the describe definition of the e2e tests like Default Backend The complete list of tests can be found here Custom docker image \u00b6 In some cases, it can be useful to build a docker image and publish such an image to a private or custom registry location. This can be done setting two environment variables, REGISTRY and TAG export TAG=\"dev\" export REGISTRY=\"$USER\" make build image and then publish such version with docker push $REGISTRY/controller:$TAG","title":"Development"},{"location":"development/#developing-for-nginx-ingress-controller","text":"This document explains how to get started with developing for NGINX Ingress controller.","title":"Developing for NGINX Ingress Controller"},{"location":"development/#prerequisites","text":"Install Go 1.14 or later. Note The project uses Go Modules Install Docker Important The majority of make tasks run as docker containers","title":"Prerequisites"},{"location":"development/#quick-start","text":"Fork the repository Clone the repository to any location in your work station Add a GO111MODULE environment variable with export GO111MODULE=on Run go mod download to install dependencies","title":"Quick Start"},{"location":"development/#local-build","text":"Start a local Kubernetes cluster using kind , build and deploy the ingress controller make dev-env","title":"Local build"},{"location":"development/#testing","text":"Run go unit tests make test Run unit-tests for lua code make lua-test Lua tests are located in the directory rootfs/etc/nginx/lua/test Important Test files must follow the naming convention _test.lua or it will be ignored Run e2e test suite make kind-e2e-test To limit the scope of the tests to execute, we can use the environment variable FOCUS FOCUS=\"no-auth-locations\" make kind-e2e-test Note The variable FOCUS defines Ginkgo Focused Specs Valid values are defined in the describe definition of the e2e tests like Default Backend The complete list of tests can be found here","title":"Testing"},{"location":"development/#custom-docker-image","text":"In some cases, it can be useful to build a docker image and publish such an image to a private or custom registry location. This can be done setting two environment variables, REGISTRY and TAG export TAG=\"dev\" export REGISTRY=\"$USER\" make build image and then publish such version with docker push $REGISTRY/controller:$TAG","title":"Custom docker image"},{"location":"e2e-tests/","text":"e2e test suite for NGINX Ingress Controller \u00b6 [Default Backend] change default settings \u00b6 should apply the annotation to the default backend [Default Backend] \u00b6 should return 404 sending requests when only a default backend is running enables access logging for default backend disables access logging for default backend [Default Backend] custom service \u00b6 uses custom default backend that returns 200 as status code [Default Backend] SSL \u00b6 should return a self generated SSL certificate [TCP] tcp-services \u00b6 should expose a TCP service should expose an ExternalName TCP service auth-* \u00b6 should return status code 200 when no authentication is configured should return status code 503 when authentication is configured with an invalid secret should return status code 401 when authentication is configured but Authorization header is not configured should return status code 401 when authentication is configured and Authorization header is sent with invalid credentials should return status code 200 when authentication is configured and Authorization header is sent should return status code 200 when authentication is configured with a map and Authorization header is sent should return status code 401 when authentication is configured with invalid content and Authorization header is sent proxy_set_header My-Custom-Header 42; proxy_set_header My-Custom-Header 42; proxy_set_header 'My-Custom-Header' '42'; retains cookie set by external authentication server should return status code 200 when signed in should redirect to signin url when not signed in keeps processing new ingresses even if one of the existing ingresses is misconfigured should return status code 200 when signed in after auth backend is deleted should deny login for different location on same server should deny login for different servers should redirect to signin url when not signed in affinitymode \u00b6 Balanced affinity mode should balance Check persistent affinity mode proxy-* \u00b6 should set proxy_redirect to off should set proxy_redirect to default should set proxy_redirect to hello.com goodbye.com should set proxy client-max-body-size to 8m should not set proxy client-max-body-size to incorrect value should set valid proxy timeouts should not set invalid proxy timeouts should turn on proxy-buffering should turn off proxy-request-buffering should build proxy next upstream should setup proxy cookies should change the default proxy HTTP version affinity session-cookie-name \u00b6 should set sticky cookie SERVERID should change cookie name on ingress definition change should set the path to /something on the generated cookie does not set the path to / on the generated cookie if there's more than one rule referring to the same backend should set cookie with expires should work with use-regex annotation and session-cookie-path should warn user when use-regex is true and session-cookie-path is not set should not set affinity across all server locations when using separate ingresses should set sticky cookie without host mirror-* \u00b6 should set mirror-target to http://localhost/mirror should set mirror-target to https://test.env.com/$request_uri should disable mirror-request-body canary-* \u00b6 should response with a 200 status from the mainline upstream when requests are made to the mainline ingress should return 404 status for requests to the canary if no matching ingress is found should return the correct status codes when endpoints are unavailable should route requests to the correct upstream if mainline ingress is created before the canary ingress should route requests to the correct upstream if mainline ingress is created after the canary ingress should route requests to the correct upstream if the mainline ingress is modified should route requests to the correct upstream if the canary ingress is modified should route requests to the correct upstream should route requests to the correct upstream should route requests to the correct upstream should route requests to the correct upstream should routes to mainline upstream when the given Regex causes error should route requests to the correct upstream should route requests to the correct upstream should route requests to the correct upstream should not use canary as a catch-all server should not use canary with domain as a server does not crash when canary ingress has multiple paths to the same non-matching backend limit-rate \u00b6 Check limit-rate annotation force-ssl-redirect \u00b6 should redirect to https http2-push-preload \u00b6 enable the http2-push-preload directive proxy-ssl-* \u00b6 should set valid proxy-ssl-secret should set valid proxy-ssl-secret, proxy-ssl-verify to on, and proxy-ssl-verify-depth to 2 should set valid proxy-ssl-secret, proxy-ssl-ciphers to HIGH:!AES should set valid proxy-ssl-secret, proxy-ssl-protocols proxy-ssl-location-only flag should change the nginx config server part modsecurity owasp \u00b6 should enable modsecurity should enable modsecurity with transaction ID and OWASP rules should disable modsecurity should enable modsecurity with snippet should enable modsecurity without using 'modsecurity on;' should disable modsecurity using 'modsecurity off;' should enable modsecurity with snippet and block requests should enable modsecurity globally and with modsecurity-snippet block requests backend-protocol - GRPC \u00b6 should use grpc_pass in the configuration file should return OK for service with backend protocol GRPC should return OK for service with backend protocol GRPCS cors-* \u00b6 should enable cors should set cors methods to only allow POST, GET should set cors max-age should disable cors allow credentials should allow origin for cors should allow headers for cors influxdb-* \u00b6 should send the request metric to the influxdb server Annotation - limit-connections \u00b6 should limit-connections client-body-buffer-size \u00b6 should set client_body_buffer_size to 1000 should set client_body_buffer_size to 1K should set client_body_buffer_size to 1k should set client_body_buffer_size to 1m should set client_body_buffer_size to 1M should not set client_body_buffer_size to invalid 1b default-backend \u00b6 should use a custom default backend as upstream connection-proxy-header \u00b6 set connection header to keep-alive upstream-vhost \u00b6 set host to upstreamvhost.bar.com custom-http-errors \u00b6 configures Nginx correctly server-snippet \u00b6 rewrite-target use-regex enable-rewrite-log \u00b6 should write rewrite logs should use correct longest path match should use ~* location modifier if regex annotation is present should fail to use longest match for documented warning should allow for custom rewrite parameters app-root \u00b6 should redirect to /foo whitelist-source-range \u00b6 should set valid ip whitelist range enable-access-log enable-rewrite-log \u00b6 set access_log off set rewrite_log on x-forwarded-prefix \u00b6 should set the X-Forwarded-Prefix to the annotation value should not add X-Forwarded-Prefix if the annotation value is empty configuration-snippet \u00b6 in all locations backend-protocol - FastCGI \u00b6 should use fastcgi_pass in the configuration file should add fastcgi_index in the configuration file should add fastcgi_param in the configuration file should return OK for service with backend protocol FastCGI from-to-www-redirect \u00b6 should redirect from www HTTP to HTTP should redirect from www HTTPS to HTTPS permanen-redirect permanen-redirect-code \u00b6 should respond with a standard redirect code should respond with a custom redirect code upstream-hash-by-* \u00b6 should connect to the same pod should connect to the same subset of pods backend-protocol \u00b6 should set backend protocol to https:// and use proxy_pass should set backend protocol to grpc:// and use grpc_pass should set backend protocol to grpcs:// and use grpc_pass should set backend protocol to '' and use fastcgi_pass should set backend protocol to '' and use ajp_pass satisfy \u00b6 should configure satisfy directive correctly should allow multiple auth with satisfy any server-alias \u00b6 should return status code 200 for host 'foo' and 404 for 'bar' should return status code 200 for host 'foo' and 'bar' should return status code 200 for hosts defined in two ingresses, different path with one alias ssl-ciphers \u00b6 should change ssl ciphers auth-tls-* \u00b6 should set valid auth-tls-secret should set valid auth-tls-secret, sslVerify to off, and sslVerifyDepth to 2 should set valid auth-tls-secret, pass certificate to upstream, and error page should validate auth-tls-verify-client [Status] status update \u00b6 should update status field after client-go reconnection Debug CLI \u00b6 should list the backend servers should get information for a specific backend server should produce valid JSON for /dbg general [Memory Leak] Dynamic Certificates \u00b6 should not leak memory from ingress SSL certificates or configuration updates single ingress - multiple hosts \u00b6 should set the correct $service_name NGINX variable [Ingress] [PathType] exact \u00b6 should choose exact location for /exact [Security] request smuggling \u00b6 should not return body content from error_page [SSL] [Flag] default-ssl-certificate \u00b6 uses default ssl certificate for catch-all ingress uses default ssl certificate for host based ingress when configured certificate does not match host access-log \u00b6 use the default configuration use the specified configuration use the specified configuration use the specified configuration use the specified configuration [Lua] lua-shared-dicts \u00b6 configures lua shared dicts server-tokens \u00b6 should not exists Server header in the response should exists Server header in the response when is enabled use-proxy-protocol \u00b6 should respect port passed by the PROXY Protocol should respect proto passed by the PROXY Protocol server port should enable PROXY Protocol for HTTPS [Flag] custom HTTP and HTTPS ports \u00b6 should set X-Forwarded-Port headers accordingly when listening on a non-default HTTP port should set X-Forwarded-Port header to 443 should set the X-Forwarded-Port header to 443 [Security] no-auth-locations \u00b6 should return status code 401 when accessing '/' unauthentication should return status code 200 when accessing '/' authentication should return status code 200 when accessing '/noauth' unauthenticated Dynamic $proxy_host \u00b6 should exist a proxy_host should exist a proxy_host using the upstream-vhost annotation value proxy-connect-timeout \u00b6 should set valid proxy timeouts using configmap values should not set invalid proxy timeouts using configmap values [Security] Pod Security Policies \u00b6 should be running with a Pod Security Policy Geoip2 \u00b6 should only allow requests from specific countries [Security] Pod Security Policies with volumes \u00b6 should be running with a Pod Security Policy enable-multi-accept \u00b6 should be enabled by default should be enabled when set to true should be disabled when set to false log-format-* \u00b6 should disable the log-format-escape-json by default should enable the log-format-escape-json should disable the log-format-escape-json log-format-escape-json enabled log-format-escape-json disabled [Flag] ingress-class \u00b6 should ignore Ingress with class should ignore Ingress with no class should delete Ingress when class is removed check scenarios for IngressClass and ingress.class annotation ssl-ciphers \u00b6 Add ssl ciphers proxy-next-upstream \u00b6 should build proxy next upstream using configmap values [Security] global-auth-url \u00b6 should return status code 401 when request any protected service should return status code 200 when request whitelisted (via no-auth-locations) service and 401 when request protected service should return status code 200 when request whitelisted (via ingress annotation) service and 401 when request protected service should still return status code 200 after auth backend is deleted using cache [Security] block-* \u00b6 should block CIDRs defined in the ConfigMap should block User-Agents defined in the ConfigMap should block Referers defined in the ConfigMap plugins \u00b6 should exist a x-hello-world header Configmap - limit-rate \u00b6 Check limit-rate config use-forwarded-headers \u00b6 should trust X-Forwarded headers when setting is true should not trust X-Forwarded headers when setting is false proxy-send-timeout \u00b6 should set valid proxy send timeouts using configmap values should not set invalid proxy send timeouts using configmap values Add no tls redirect locations \u00b6 Check no tls redirect locations config add-headers \u00b6 Add a custom header Add multiple custom headers hash size \u00b6 should set server_names_hash_bucket_size should set server_names_hash_max_size should set proxy-headers-hash-bucket-size should set proxy-headers-hash-max-size should set variables-hash-bucket-size should set variables-hash-max-size should set vmap-hash-bucket-size keep-alive keep-alive-requests \u00b6 should set keepalive_timeout should set keepalive_requests should set keepalive connection to upstream server should set keep alive connection timeout to upstream server should set the request count to upstream server through one keep alive connection [Flag] disable-catch-all \u00b6 should ignore catch all Ingress should delete Ingress updated to catch-all should allow Ingress with both a default backend and rules main-snippet \u00b6 should add value of main-snippet setting to nginx config [SSL] TLS protocols, ciphers and headers) \u00b6 setting cipher suite enforcing TLS v1.0 setting max-age parameter setting includeSubDomains parameter setting preload parameter overriding what's set from the upstream should not use ports during the HTTP to HTTPS redirection should not use ports or X-Forwarded-Host during the HTTP to HTTPS redirection Configmap change \u00b6 should reload after an update in the configuration proxy-read-timeout \u00b6 should set valid proxy read timeouts using configmap values should not set invalid proxy read timeouts using configmap values [Security] modsecurity-snippet \u00b6 should add value of modsecurity-snippet setting to nginx config OCSP \u00b6 should enable OCSP and contain stapling information in the connection reuse-port \u00b6 reuse port should be enabled by default reuse port should be disabled reuse port should be enabled [Shutdown] Graceful shutdown with pending request \u00b6 should let slow requests finish before shutting down [Shutdown] ingress controller \u00b6 should shutdown in less than 60 secons without pending connections should shutdown after waiting 60 seconds for pending connections to be closed should shutdown after waiting 150 seconds for pending connections to be closed [Service] backend status code 503 \u00b6 should return 503 when backend service does not exist should return 503 when all backend service endpoints are unavailable [Service] Type ExternalName \u00b6 works with external name set to incomplete fdqn should return 200 for service type=ExternalName without a port defined should return 200 for service type=ExternalName with a port defined should return status 502 for service type=ExternalName with an invalid host should return 200 for service type=ExternalName using a port name should update the external name after a service update","title":"e2e test suite for [NGINX Ingress Controller](https://github.com/kubernetes/ingress-nginx/tree/master/)"},{"location":"e2e-tests/#e2e-test-suite-for-nginx-ingress-controller","text":"","title":"e2e test suite for NGINX Ingress Controller"},{"location":"e2e-tests/#default-backend-change-default-settings","text":"should apply the annotation to the default backend","title":"[Default Backend] change default settings"},{"location":"e2e-tests/#default-backend","text":"should return 404 sending requests when only a default backend is running enables access logging for default backend disables access logging for default backend","title":"[Default Backend]"},{"location":"e2e-tests/#default-backend-custom-service","text":"uses custom default backend that returns 200 as status code","title":"[Default Backend] custom service"},{"location":"e2e-tests/#default-backend-ssl","text":"should return a self generated SSL certificate","title":"[Default Backend] SSL"},{"location":"e2e-tests/#tcp-tcp-services","text":"should expose a TCP service should expose an ExternalName TCP service","title":"[TCP] tcp-services"},{"location":"e2e-tests/#auth-","text":"should return status code 200 when no authentication is configured should return status code 503 when authentication is configured with an invalid secret should return status code 401 when authentication is configured but Authorization header is not configured should return status code 401 when authentication is configured and Authorization header is sent with invalid credentials should return status code 200 when authentication is configured and Authorization header is sent should return status code 200 when authentication is configured with a map and Authorization header is sent should return status code 401 when authentication is configured with invalid content and Authorization header is sent proxy_set_header My-Custom-Header 42; proxy_set_header My-Custom-Header 42; proxy_set_header 'My-Custom-Header' '42'; retains cookie set by external authentication server should return status code 200 when signed in should redirect to signin url when not signed in keeps processing new ingresses even if one of the existing ingresses is misconfigured should return status code 200 when signed in after auth backend is deleted should deny login for different location on same server should deny login for different servers should redirect to signin url when not signed in","title":"auth-*"},{"location":"e2e-tests/#affinitymode","text":"Balanced affinity mode should balance Check persistent affinity mode","title":"affinitymode"},{"location":"e2e-tests/#proxy-","text":"should set proxy_redirect to off should set proxy_redirect to default should set proxy_redirect to hello.com goodbye.com should set proxy client-max-body-size to 8m should not set proxy client-max-body-size to incorrect value should set valid proxy timeouts should not set invalid proxy timeouts should turn on proxy-buffering should turn off proxy-request-buffering should build proxy next upstream should setup proxy cookies should change the default proxy HTTP version","title":"proxy-*"},{"location":"e2e-tests/#affinity-session-cookie-name","text":"should set sticky cookie SERVERID should change cookie name on ingress definition change should set the path to /something on the generated cookie does not set the path to / on the generated cookie if there's more than one rule referring to the same backend should set cookie with expires should work with use-regex annotation and session-cookie-path should warn user when use-regex is true and session-cookie-path is not set should not set affinity across all server locations when using separate ingresses should set sticky cookie without host","title":"affinity session-cookie-name"},{"location":"e2e-tests/#mirror-","text":"should set mirror-target to http://localhost/mirror should set mirror-target to https://test.env.com/$request_uri should disable mirror-request-body","title":"mirror-*"},{"location":"e2e-tests/#canary-","text":"should response with a 200 status from the mainline upstream when requests are made to the mainline ingress should return 404 status for requests to the canary if no matching ingress is found should return the correct status codes when endpoints are unavailable should route requests to the correct upstream if mainline ingress is created before the canary ingress should route requests to the correct upstream if mainline ingress is created after the canary ingress should route requests to the correct upstream if the mainline ingress is modified should route requests to the correct upstream if the canary ingress is modified should route requests to the correct upstream should route requests to the correct upstream should route requests to the correct upstream should route requests to the correct upstream should routes to mainline upstream when the given Regex causes error should route requests to the correct upstream should route requests to the correct upstream should route requests to the correct upstream should not use canary as a catch-all server should not use canary with domain as a server does not crash when canary ingress has multiple paths to the same non-matching backend","title":"canary-*"},{"location":"e2e-tests/#limit-rate","text":"Check limit-rate annotation","title":"limit-rate"},{"location":"e2e-tests/#force-ssl-redirect","text":"should redirect to https","title":"force-ssl-redirect"},{"location":"e2e-tests/#http2-push-preload","text":"enable the http2-push-preload directive","title":"http2-push-preload"},{"location":"e2e-tests/#proxy-ssl-","text":"should set valid proxy-ssl-secret should set valid proxy-ssl-secret, proxy-ssl-verify to on, and proxy-ssl-verify-depth to 2 should set valid proxy-ssl-secret, proxy-ssl-ciphers to HIGH:!AES should set valid proxy-ssl-secret, proxy-ssl-protocols proxy-ssl-location-only flag should change the nginx config server part","title":"proxy-ssl-*"},{"location":"e2e-tests/#modsecurity-owasp","text":"should enable modsecurity should enable modsecurity with transaction ID and OWASP rules should disable modsecurity should enable modsecurity with snippet should enable modsecurity without using 'modsecurity on;' should disable modsecurity using 'modsecurity off;' should enable modsecurity with snippet and block requests should enable modsecurity globally and with modsecurity-snippet block requests","title":"modsecurity owasp"},{"location":"e2e-tests/#backend-protocol-grpc","text":"should use grpc_pass in the configuration file should return OK for service with backend protocol GRPC should return OK for service with backend protocol GRPCS","title":"backend-protocol - GRPC"},{"location":"e2e-tests/#cors-","text":"should enable cors should set cors methods to only allow POST, GET should set cors max-age should disable cors allow credentials should allow origin for cors should allow headers for cors","title":"cors-*"},{"location":"e2e-tests/#influxdb-","text":"should send the request metric to the influxdb server","title":"influxdb-*"},{"location":"e2e-tests/#annotation-limit-connections","text":"should limit-connections","title":"Annotation - limit-connections"},{"location":"e2e-tests/#client-body-buffer-size","text":"should set client_body_buffer_size to 1000 should set client_body_buffer_size to 1K should set client_body_buffer_size to 1k should set client_body_buffer_size to 1m should set client_body_buffer_size to 1M should not set client_body_buffer_size to invalid 1b","title":"client-body-buffer-size"},{"location":"e2e-tests/#default-backend_1","text":"should use a custom default backend as upstream","title":"default-backend"},{"location":"e2e-tests/#connection-proxy-header","text":"set connection header to keep-alive","title":"connection-proxy-header"},{"location":"e2e-tests/#upstream-vhost","text":"set host to upstreamvhost.bar.com","title":"upstream-vhost"},{"location":"e2e-tests/#custom-http-errors","text":"configures Nginx correctly","title":"custom-http-errors"},{"location":"e2e-tests/#server-snippet","text":"","title":"server-snippet"},{"location":"e2e-tests/#rewrite-target-use-regex-enable-rewrite-log","text":"should write rewrite logs should use correct longest path match should use ~* location modifier if regex annotation is present should fail to use longest match for documented warning should allow for custom rewrite parameters","title":"rewrite-target use-regex enable-rewrite-log"},{"location":"e2e-tests/#app-root","text":"should redirect to /foo","title":"app-root"},{"location":"e2e-tests/#whitelist-source-range","text":"should set valid ip whitelist range","title":"whitelist-source-range"},{"location":"e2e-tests/#enable-access-log-enable-rewrite-log","text":"set access_log off set rewrite_log on","title":"enable-access-log enable-rewrite-log"},{"location":"e2e-tests/#x-forwarded-prefix","text":"should set the X-Forwarded-Prefix to the annotation value should not add X-Forwarded-Prefix if the annotation value is empty","title":"x-forwarded-prefix"},{"location":"e2e-tests/#configuration-snippet","text":"in all locations","title":"configuration-snippet"},{"location":"e2e-tests/#backend-protocol-fastcgi","text":"should use fastcgi_pass in the configuration file should add fastcgi_index in the configuration file should add fastcgi_param in the configuration file should return OK for service with backend protocol FastCGI","title":"backend-protocol - FastCGI"},{"location":"e2e-tests/#from-to-www-redirect","text":"should redirect from www HTTP to HTTP should redirect from www HTTPS to HTTPS","title":"from-to-www-redirect"},{"location":"e2e-tests/#permanen-redirect-permanen-redirect-code","text":"should respond with a standard redirect code should respond with a custom redirect code","title":"permanen-redirect permanen-redirect-code"},{"location":"e2e-tests/#upstream-hash-by-","text":"should connect to the same pod should connect to the same subset of pods","title":"upstream-hash-by-*"},{"location":"e2e-tests/#backend-protocol","text":"should set backend protocol to https:// and use proxy_pass should set backend protocol to grpc:// and use grpc_pass should set backend protocol to grpcs:// and use grpc_pass should set backend protocol to '' and use fastcgi_pass should set backend protocol to '' and use ajp_pass","title":"backend-protocol"},{"location":"e2e-tests/#satisfy","text":"should configure satisfy directive correctly should allow multiple auth with satisfy any","title":"satisfy"},{"location":"e2e-tests/#server-alias","text":"should return status code 200 for host 'foo' and 404 for 'bar' should return status code 200 for host 'foo' and 'bar' should return status code 200 for hosts defined in two ingresses, different path with one alias","title":"server-alias"},{"location":"e2e-tests/#ssl-ciphers","text":"should change ssl ciphers","title":"ssl-ciphers"},{"location":"e2e-tests/#auth-tls-","text":"should set valid auth-tls-secret should set valid auth-tls-secret, sslVerify to off, and sslVerifyDepth to 2 should set valid auth-tls-secret, pass certificate to upstream, and error page should validate auth-tls-verify-client","title":"auth-tls-*"},{"location":"e2e-tests/#status-status-update","text":"should update status field after client-go reconnection","title":"[Status] status update"},{"location":"e2e-tests/#debug-cli","text":"should list the backend servers should get information for a specific backend server should produce valid JSON for /dbg general","title":"Debug CLI"},{"location":"e2e-tests/#memory-leak-dynamic-certificates","text":"should not leak memory from ingress SSL certificates or configuration updates","title":"[Memory Leak] Dynamic Certificates"},{"location":"e2e-tests/#single-ingress-multiple-hosts","text":"should set the correct $service_name NGINX variable","title":"single ingress - multiple hosts"},{"location":"e2e-tests/#ingress-pathtype-exact","text":"should choose exact location for /exact","title":"[Ingress] [PathType] exact"},{"location":"e2e-tests/#security-request-smuggling","text":"should not return body content from error_page","title":"[Security] request smuggling"},{"location":"e2e-tests/#ssl-flag-default-ssl-certificate","text":"uses default ssl certificate for catch-all ingress uses default ssl certificate for host based ingress when configured certificate does not match host","title":"[SSL] [Flag] default-ssl-certificate"},{"location":"e2e-tests/#access-log","text":"use the default configuration use the specified configuration use the specified configuration use the specified configuration use the specified configuration","title":"access-log"},{"location":"e2e-tests/#lua-lua-shared-dicts","text":"configures lua shared dicts","title":"[Lua] lua-shared-dicts"},{"location":"e2e-tests/#server-tokens","text":"should not exists Server header in the response should exists Server header in the response when is enabled","title":"server-tokens"},{"location":"e2e-tests/#use-proxy-protocol","text":"should respect port passed by the PROXY Protocol should respect proto passed by the PROXY Protocol server port should enable PROXY Protocol for HTTPS","title":"use-proxy-protocol"},{"location":"e2e-tests/#flag-custom-http-and-https-ports","text":"should set X-Forwarded-Port headers accordingly when listening on a non-default HTTP port should set X-Forwarded-Port header to 443 should set the X-Forwarded-Port header to 443","title":"[Flag] custom HTTP and HTTPS ports"},{"location":"e2e-tests/#security-no-auth-locations","text":"should return status code 401 when accessing '/' unauthentication should return status code 200 when accessing '/' authentication should return status code 200 when accessing '/noauth' unauthenticated","title":"[Security] no-auth-locations"},{"location":"e2e-tests/#dynamic-proxy_host","text":"should exist a proxy_host should exist a proxy_host using the upstream-vhost annotation value","title":"Dynamic $proxy_host"},{"location":"e2e-tests/#proxy-connect-timeout","text":"should set valid proxy timeouts using configmap values should not set invalid proxy timeouts using configmap values","title":"proxy-connect-timeout"},{"location":"e2e-tests/#security-pod-security-policies","text":"should be running with a Pod Security Policy","title":"[Security] Pod Security Policies"},{"location":"e2e-tests/#geoip2","text":"should only allow requests from specific countries","title":"Geoip2"},{"location":"e2e-tests/#security-pod-security-policies-with-volumes","text":"should be running with a Pod Security Policy","title":"[Security] Pod Security Policies with volumes"},{"location":"e2e-tests/#enable-multi-accept","text":"should be enabled by default should be enabled when set to true should be disabled when set to false","title":"enable-multi-accept"},{"location":"e2e-tests/#log-format-","text":"should disable the log-format-escape-json by default should enable the log-format-escape-json should disable the log-format-escape-json log-format-escape-json enabled log-format-escape-json disabled","title":"log-format-*"},{"location":"e2e-tests/#flag-ingress-class","text":"should ignore Ingress with class should ignore Ingress with no class should delete Ingress when class is removed check scenarios for IngressClass and ingress.class annotation","title":"[Flag] ingress-class"},{"location":"e2e-tests/#ssl-ciphers_1","text":"Add ssl ciphers","title":"ssl-ciphers"},{"location":"e2e-tests/#proxy-next-upstream","text":"should build proxy next upstream using configmap values","title":"proxy-next-upstream"},{"location":"e2e-tests/#security-global-auth-url","text":"should return status code 401 when request any protected service should return status code 200 when request whitelisted (via no-auth-locations) service and 401 when request protected service should return status code 200 when request whitelisted (via ingress annotation) service and 401 when request protected service should still return status code 200 after auth backend is deleted using cache","title":"[Security] global-auth-url"},{"location":"e2e-tests/#security-block-","text":"should block CIDRs defined in the ConfigMap should block User-Agents defined in the ConfigMap should block Referers defined in the ConfigMap","title":"[Security] block-*"},{"location":"e2e-tests/#plugins","text":"should exist a x-hello-world header","title":"plugins"},{"location":"e2e-tests/#configmap-limit-rate","text":"Check limit-rate config","title":"Configmap - limit-rate"},{"location":"e2e-tests/#use-forwarded-headers","text":"should trust X-Forwarded headers when setting is true should not trust X-Forwarded headers when setting is false","title":"use-forwarded-headers"},{"location":"e2e-tests/#proxy-send-timeout","text":"should set valid proxy send timeouts using configmap values should not set invalid proxy send timeouts using configmap values","title":"proxy-send-timeout"},{"location":"e2e-tests/#add-no-tls-redirect-locations","text":"Check no tls redirect locations config","title":"Add no tls redirect locations"},{"location":"e2e-tests/#add-headers","text":"Add a custom header Add multiple custom headers","title":"add-headers"},{"location":"e2e-tests/#hash-size","text":"should set server_names_hash_bucket_size should set server_names_hash_max_size should set proxy-headers-hash-bucket-size should set proxy-headers-hash-max-size should set variables-hash-bucket-size should set variables-hash-max-size should set vmap-hash-bucket-size","title":"hash size"},{"location":"e2e-tests/#keep-alive-keep-alive-requests","text":"should set keepalive_timeout should set keepalive_requests should set keepalive connection to upstream server should set keep alive connection timeout to upstream server should set the request count to upstream server through one keep alive connection","title":"keep-alive keep-alive-requests"},{"location":"e2e-tests/#flag-disable-catch-all","text":"should ignore catch all Ingress should delete Ingress updated to catch-all should allow Ingress with both a default backend and rules","title":"[Flag] disable-catch-all"},{"location":"e2e-tests/#main-snippet","text":"should add value of main-snippet setting to nginx config","title":"main-snippet"},{"location":"e2e-tests/#ssl-tls-protocols-ciphers-and-headers","text":"setting cipher suite enforcing TLS v1.0 setting max-age parameter setting includeSubDomains parameter setting preload parameter overriding what's set from the upstream should not use ports during the HTTP to HTTPS redirection should not use ports or X-Forwarded-Host during the HTTP to HTTPS redirection","title":"[SSL] TLS protocols, ciphers and headers)"},{"location":"e2e-tests/#configmap-change","text":"should reload after an update in the configuration","title":"Configmap change"},{"location":"e2e-tests/#proxy-read-timeout","text":"should set valid proxy read timeouts using configmap values should not set invalid proxy read timeouts using configmap values","title":"proxy-read-timeout"},{"location":"e2e-tests/#security-modsecurity-snippet","text":"should add value of modsecurity-snippet setting to nginx config","title":"[Security] modsecurity-snippet"},{"location":"e2e-tests/#ocsp","text":"should enable OCSP and contain stapling information in the connection","title":"OCSP"},{"location":"e2e-tests/#reuse-port","text":"reuse port should be enabled by default reuse port should be disabled reuse port should be enabled","title":"reuse-port"},{"location":"e2e-tests/#shutdown-graceful-shutdown-with-pending-request","text":"should let slow requests finish before shutting down","title":"[Shutdown] Graceful shutdown with pending request"},{"location":"e2e-tests/#shutdown-ingress-controller","text":"should shutdown in less than 60 secons without pending connections should shutdown after waiting 60 seconds for pending connections to be closed should shutdown after waiting 150 seconds for pending connections to be closed","title":"[Shutdown] ingress controller"},{"location":"e2e-tests/#service-backend-status-code-503","text":"should return 503 when backend service does not exist should return 503 when all backend service endpoints are unavailable","title":"[Service] backend status code 503"},{"location":"e2e-tests/#service-type-externalname","text":"works with external name set to incomplete fdqn should return 200 for service type=ExternalName without a port defined should return 200 for service type=ExternalName with a port defined should return status 502 for service type=ExternalName with an invalid host should return 200 for service type=ExternalName using a port name should update the external name after a service update","title":"[Service] Type ExternalName"},{"location":"how-it-works/","text":"How it works \u00b6 The objective of this document is to explain how the NGINX Ingress controller works, in particular how the NGINX model is built and why we need one. NGINX configuration \u00b6 The goal of this Ingress controller is the assembly of a configuration file (nginx.conf). The main implication of this requirement is the need to reload NGINX after any change in the configuration file. Though it is important to note that we don't reload Nginx on changes that impact only an upstream configuration (i.e Endpoints change when you deploy your app) . We use lua-nginx-module to achieve this. Check below to learn more about how it's done. NGINX model \u00b6 Usually, a Kubernetes Controller utilizes the synchronization loop pattern to check if the desired state in the controller is updated or a change is required. To this purpose, we need to build a model using different objects from the cluster, in particular (in no special order) Ingresses, Services, Endpoints, Secrets, and Configmaps to generate a point in time configuration file that reflects the state of the cluster. To get this object from the cluster, we use Kubernetes Informers , in particular, FilteredSharedInformer . This informers allows reacting to changes in using callbacks to individual changes when a new object is added, modified or removed. Unfortunately, there is no way to know if a particular change is going to affect the final configuration file. Therefore on every change, we have to rebuild a new model from scratch based on the state of cluster and compare it to the current model. If the new model equals to the current one, then we avoid generating a new NGINX configuration and triggering a reload. Otherwise, we check if the difference is only about Endpoints. If so we then send the new list of Endpoints to a Lua handler running inside Nginx using HTTP POST request and again avoid generating a new NGINX configuration and triggering a reload. If the difference between running and new model is about more than just Endpoints we create a new NGINX configuration based on the new model, replace the current model and trigger a reload. One of the uses of the model is to avoid unnecessary reloads when there's no change in the state and to detect conflicts in definitions. The final representation of the NGINX configuration is generated from a Go template using the new model as input for the variables required by the template. Building the NGINX model \u00b6 Building a model is an expensive operation, for this reason, the use of the synchronization loop is a must. By using a work queue it is possible to not lose changes and remove the use of sync.Mutex to force a single execution of the sync loop and additionally it is possible to create a time window between the start and end of the sync loop that allows us to discard unnecessary updates. It is important to understand that any change in the cluster could generate events that the informer will send to the controller and one of the reasons for the work queue . Operations to build the model: Order Ingress rules by CreationTimestamp field, i.e., old rules first. If the same path for the same host is defined in more than one Ingress, the oldest rule wins. If more than one Ingress contains a TLS section for the same host, the oldest rule wins. If multiple Ingresses define an annotation that affects the configuration of the Server block, the oldest rule wins. Create a list of NGINX Servers (per hostname) Create a list of NGINX Upstreams If multiple Ingresses define different paths for the same host, the ingress controller will merge the definitions. Annotations are applied to all the paths in the Ingress. Multiple Ingresses can define different annotations. These definitions are not shared between Ingresses. When a reload is required \u00b6 The next list describes the scenarios when a reload is required: New Ingress Resource Created. TLS section is added to existing Ingress. Change in Ingress annotations that impacts more than just upstream configuration. For instance load-balance annotation does not require a reload. A path is added/removed from an Ingress. An Ingress, Service, Secret is removed. Some missing referenced object from the Ingress is available, like a Service or Secret. A Secret is updated. Avoiding reloads \u00b6 In some cases, it is possible to avoid reloads, in particular when there is a change in the endpoints, i.e., a pod is started or replaced. It is out of the scope of this Ingress controller to remove reloads completely. This would require an incredible amount of work and at some point makes no sense. This can change only if NGINX changes the way new configurations are read, basically, new changes do not replace worker processes. Avoiding reloads on Endpoints changes \u00b6 On every endpoint change the controller fetches endpoints from all the services it sees and generates corresponding Backend objects. It then sends these objects to a Lua handler running inside Nginx. The Lua code in turn stores those backends in a shared memory zone. Then for every request Lua code running in balancer_by_lua context detects what endpoints it should choose upstream peer from and applies the configured load balancing algorithm to choose the peer. Then Nginx takes care of the rest. This way we avoid reloading Nginx on endpoint changes. Note that this includes annotation changes that affects only upstream configuration in Nginx as well. In a relatively big clusters with frequently deploying apps this feature saves significant number of Nginx reloads which can otherwise affect response latency, load balancing quality (after every reload Nginx resets the state of load balancing) and so on. Avoiding outage from wrong configuration \u00b6 Because the ingress controller works using the synchronization loop pattern , it is applying the configuration for all matching objects. In case some Ingress objects have a broken configuration, for example a syntax error in the nginx.ingress.kubernetes.io/configuration-snippet annotation, the generated configuration becomes invalid, does not reload and hence no more ingresses will be taken into account. To prevent this situation to happen, the nginx ingress controller optionally exposes a validating admission webhook server to ensure the validity of incoming ingress objects. This webhook appends the incoming ingress objects to the list of ingresses, generates the configuration and calls nginx to ensure the configuration has no syntax errors.","title":"How it works"},{"location":"how-it-works/#how-it-works","text":"The objective of this document is to explain how the NGINX Ingress controller works, in particular how the NGINX model is built and why we need one.","title":"How it works"},{"location":"how-it-works/#nginx-configuration","text":"The goal of this Ingress controller is the assembly of a configuration file (nginx.conf). The main implication of this requirement is the need to reload NGINX after any change in the configuration file. Though it is important to note that we don't reload Nginx on changes that impact only an upstream configuration (i.e Endpoints change when you deploy your app) . We use lua-nginx-module to achieve this. Check below to learn more about how it's done.","title":"NGINX configuration"},{"location":"how-it-works/#nginx-model","text":"Usually, a Kubernetes Controller utilizes the synchronization loop pattern to check if the desired state in the controller is updated or a change is required. To this purpose, we need to build a model using different objects from the cluster, in particular (in no special order) Ingresses, Services, Endpoints, Secrets, and Configmaps to generate a point in time configuration file that reflects the state of the cluster. To get this object from the cluster, we use Kubernetes Informers , in particular, FilteredSharedInformer . This informers allows reacting to changes in using callbacks to individual changes when a new object is added, modified or removed. Unfortunately, there is no way to know if a particular change is going to affect the final configuration file. Therefore on every change, we have to rebuild a new model from scratch based on the state of cluster and compare it to the current model. If the new model equals to the current one, then we avoid generating a new NGINX configuration and triggering a reload. Otherwise, we check if the difference is only about Endpoints. If so we then send the new list of Endpoints to a Lua handler running inside Nginx using HTTP POST request and again avoid generating a new NGINX configuration and triggering a reload. If the difference between running and new model is about more than just Endpoints we create a new NGINX configuration based on the new model, replace the current model and trigger a reload. One of the uses of the model is to avoid unnecessary reloads when there's no change in the state and to detect conflicts in definitions. The final representation of the NGINX configuration is generated from a Go template using the new model as input for the variables required by the template.","title":"NGINX model"},{"location":"how-it-works/#building-the-nginx-model","text":"Building a model is an expensive operation, for this reason, the use of the synchronization loop is a must. By using a work queue it is possible to not lose changes and remove the use of sync.Mutex to force a single execution of the sync loop and additionally it is possible to create a time window between the start and end of the sync loop that allows us to discard unnecessary updates. It is important to understand that any change in the cluster could generate events that the informer will send to the controller and one of the reasons for the work queue . Operations to build the model: Order Ingress rules by CreationTimestamp field, i.e., old rules first. If the same path for the same host is defined in more than one Ingress, the oldest rule wins. If more than one Ingress contains a TLS section for the same host, the oldest rule wins. If multiple Ingresses define an annotation that affects the configuration of the Server block, the oldest rule wins. Create a list of NGINX Servers (per hostname) Create a list of NGINX Upstreams If multiple Ingresses define different paths for the same host, the ingress controller will merge the definitions. Annotations are applied to all the paths in the Ingress. Multiple Ingresses can define different annotations. These definitions are not shared between Ingresses.","title":"Building the NGINX model"},{"location":"how-it-works/#when-a-reload-is-required","text":"The next list describes the scenarios when a reload is required: New Ingress Resource Created. TLS section is added to existing Ingress. Change in Ingress annotations that impacts more than just upstream configuration. For instance load-balance annotation does not require a reload. A path is added/removed from an Ingress. An Ingress, Service, Secret is removed. Some missing referenced object from the Ingress is available, like a Service or Secret. A Secret is updated.","title":"When a reload is required"},{"location":"how-it-works/#avoiding-reloads","text":"In some cases, it is possible to avoid reloads, in particular when there is a change in the endpoints, i.e., a pod is started or replaced. It is out of the scope of this Ingress controller to remove reloads completely. This would require an incredible amount of work and at some point makes no sense. This can change only if NGINX changes the way new configurations are read, basically, new changes do not replace worker processes.","title":"Avoiding reloads"},{"location":"how-it-works/#avoiding-reloads-on-endpoints-changes","text":"On every endpoint change the controller fetches endpoints from all the services it sees and generates corresponding Backend objects. It then sends these objects to a Lua handler running inside Nginx. The Lua code in turn stores those backends in a shared memory zone. Then for every request Lua code running in balancer_by_lua context detects what endpoints it should choose upstream peer from and applies the configured load balancing algorithm to choose the peer. Then Nginx takes care of the rest. This way we avoid reloading Nginx on endpoint changes. Note that this includes annotation changes that affects only upstream configuration in Nginx as well. In a relatively big clusters with frequently deploying apps this feature saves significant number of Nginx reloads which can otherwise affect response latency, load balancing quality (after every reload Nginx resets the state of load balancing) and so on.","title":"Avoiding reloads on Endpoints changes"},{"location":"how-it-works/#avoiding-outage-from-wrong-configuration","text":"Because the ingress controller works using the synchronization loop pattern , it is applying the configuration for all matching objects. In case some Ingress objects have a broken configuration, for example a syntax error in the nginx.ingress.kubernetes.io/configuration-snippet annotation, the generated configuration becomes invalid, does not reload and hence no more ingresses will be taken into account. To prevent this situation to happen, the nginx ingress controller optionally exposes a validating admission webhook server to ensure the validity of incoming ingress objects. This webhook appends the incoming ingress objects to the list of ingresses, generates the configuration and calls nginx to ensure the configuration has no syntax errors.","title":"Avoiding outage from wrong configuration"},{"location":"kubectl-plugin/","text":"The ingress-nginx kubectl plugin \u00b6 Installation \u00b6 Install krew , then run kubectl krew install ingress-nginx to install the plugin. Then run kubectl ingress-nginx --help to make sure the plugin is properly installed and to get a list of commands: kubectl ingress-nginx --help A kubectl plugin for inspecting your ingress-nginx deployments Usage: ingress-nginx [command] Available Commands: backends Inspect the dynamic backend information of an ingress-nginx instance certs Output the certificate data stored in an ingress-nginx pod conf Inspect the generated nginx.conf exec Execute a command inside an ingress-nginx pod general Inspect the other dynamic ingress-nginx information help Help about any command info Show information about the ingress-nginx service ingresses Provide a short summary of all of the ingress definitions lint Inspect kubernetes resources for possible issues logs Get the kubernetes logs for an ingress-nginx pod ssh ssh into a running ingress-nginx pod Flags: --as string Username to impersonate for the operation --as-group stringArray Group to impersonate for the operation, this flag can be repeated to specify multiple groups. --cache-dir string Default HTTP cache directory (default \"/Users/alexkursell/.kube/http-cache\") --certificate-authority string Path to a cert file for the certificate authority --client-certificate string Path to a client certificate file for TLS --client-key string Path to a client key file for TLS --cluster string The name of the kubeconfig cluster to use --context string The name of the kubeconfig context to use -h, --help help for ingress-nginx --insecure-skip-tls-verify If true, the server's certificate will not be checked for validity. This will make your HTTPS connections insecure --kubeconfig string Path to the kubeconfig file to use for CLI requests. -n, --namespace string If present, the namespace scope for this CLI request --request-timeout string The length of time to wait before giving up on a single server request. Non-zero values should contain a corresponding time unit (e.g. 1s, 2m, 3h). A value of zero means don't timeout requests. (default \"0\") -s, --server string The address and port of the Kubernetes API server --token string Bearer token for authentication to the API server --user string The name of the kubeconfig user to use Use \"ingress-nginx [command] --help\" for more information about a command. Common Flags \u00b6 Every subcommand supports the basic kubectl configuration flags like --namespace , --context , --client-key and so on. Subcommands that act on a particular ingress-nginx pod ( backends , certs , conf , exec , general , logs , ssh ), support the --deployment and --pod flags to select either a pod from a deployment with the given name, or a pod with the given name. The --deployment flag defaults to nginx-ingress-controller . Subcommands that inspect resources ( ingresses , lint ) support the --all-namespaces flag, which causes them to inspect resources in every namespace. Subcommands \u00b6 Note that backends , general , certs , and conf require ingress-nginx version 0.23.0 or higher. backends \u00b6 Run kubectl ingress-nginx backends to get a JSON array of the backends that an ingress-nginx controller currently knows about: $ kubectl ingress-nginx backends -n ingress-nginx [ { \"name\": \"default-apple-service-5678\", \"service\": { \"metadata\": { \"creationTimestamp\": null }, \"spec\": { \"ports\": [ { \"protocol\": \"TCP\", \"port\": 5678, \"targetPort\": 5678 } ], \"selector\": { \"app\": \"apple\" }, \"clusterIP\": \"10.97.230.121\", \"type\": \"ClusterIP\", \"sessionAffinity\": \"None\" }, \"status\": { \"loadBalancer\": {} } }, \"port\": 0, \"sslPassthrough\": false, \"endpoints\": [ { \"address\": \"10.1.3.86\", \"port\": \"5678\" } ], \"sessionAffinityConfig\": { \"name\": \"\", \"cookieSessionAffinity\": { \"name\": \"\" } }, \"upstreamHashByConfig\": { \"upstream-hash-by-subset-size\": 3 }, \"noServer\": false, \"trafficShapingPolicy\": { \"weight\": 0, \"header\": \"\", \"headerValue\": \"\", \"cookie\": \"\" } }, { \"name\": \"default-echo-service-8080\", ... }, { \"name\": \"upstream-default-backend\", ... } ] Add the --list option to show only the backend names. Add the --backend option to show only the backend with the given name. certs \u00b6 Use kubectl ingress-nginx certs --host to dump the SSL cert/key information for a given host. WARNING: This command will dump sensitive private key information. Don't blindly share the output, and certainly don't log it anywhere. $ kubectl ingress-nginx certs -n ingress-nginx --host testaddr.local -----BEGIN CERTIFICATE----- ... -----END CERTIFICATE----- -----BEGIN CERTIFICATE----- ... -----END CERTIFICATE----- -----BEGIN RSA PRIVATE KEY----- -----END RSA PRIVATE KEY----- conf \u00b6 Use kubectl ingress-nginx conf to dump the generated nginx.conf file. Add the --host option to view only the server block for that host: kubectl ingress-nginx conf -n ingress-nginx --host testaddr.local server { server_name testaddr.local ; listen 80; set $proxy_upstream_name \"-\"; set $pass_access_scheme $scheme; set $pass_server_port $server_port; set $best_http_host $http_host; set $pass_port $pass_server_port; location / { set $namespace \"\"; set $ingress_name \"\"; set $service_name \"\"; set $service_port \"0\"; set $location_path \"/\"; ... exec \u00b6 kubectl ingress-nginx exec is exactly the same as kubectl exec , with the same command flags. It will automatically choose an ingress-nginx pod to run the command in. $ kubectl ingress-nginx exec -i -n ingress-nginx -- ls /etc/nginx fastcgi_params geoip lua mime.types modsecurity modules nginx.conf opentracing.json owasp-modsecurity-crs template general \u00b6 kubectl ingress-nginx general dumps miscellaneous controller state as a JSON object. Currently it just shows the number of controller pods known to a particular controller pod. $ kubectl ingress-nginx general -n ingress-nginx { \"controllerPodsCount\": 1 } info \u00b6 Shows the internal and external IP/CNAMES for an ingress-nginx service. $ kubectl ingress-nginx info -n ingress-nginx Service cluster IP address: 10.187.253.31 LoadBalancer IP|CNAME: 35.123.123.123 Use the --service flag if your ingress-nginx LoadBalancer service is not named ingress-nginx . ingresses \u00b6 kubectl ingress-nginx ingresses , alternately kubectl ingress-nginx ing , shows a more detailed view of the ingress definitions in a namespace. Compare: $ kubectl get ingresses --all-namespaces NAMESPACE NAME HOSTS ADDRESS PORTS AGE default example-ingress1 testaddr.local,testaddr2.local localhost 80 5d default test-ingress-2 * localhost 80 5d vs $ kubectl ingress-nginx ingresses --all-namespaces NAMESPACE INGRESS NAME HOST+PATH ADDRESSES TLS SERVICE SERVICE PORT ENDPOINTS default example-ingress1 testaddr.local/etameta localhost NO pear-service 5678 5 default example-ingress1 testaddr2.local/otherpath localhost NO apple-service 5678 1 default example-ingress1 testaddr2.local/otherotherpath localhost NO pear-service 5678 5 default test-ingress-2 * localhost NO echo-service 8080 2 lint \u00b6 kubectl ingress-nginx lint can check a namespace or entire cluster for potential configuration issues. This command is especially useful when upgrading between ingress-nginx versions. $ kubectl ingress-nginx lint --all-namespaces --verbose Checking ingresses... \u2717 anamespace/this-nginx - Contains the removed session-cookie-hash annotation. Lint added for version 0.24.0 https://github.com/kubernetes/ingress-nginx/issues/3743 \u2717 othernamespace/ingress-definition-blah - The rewrite-target annotation value does not reference a capture group Lint added for version 0.22.0 https://github.com/kubernetes/ingress-nginx/issues/3174 Checking deployments... \u2717 namespace2/nginx-ingress-controller - Uses removed config flag --sort-backends Lint added for version 0.22.0 https://github.com/kubernetes/ingress-nginx/issues/3655 - Uses removed config flag --enable-dynamic-certificates Lint added for version 0.24.0 https://github.com/kubernetes/ingress-nginx/issues/3808 to show the lints added only for a particular ingress-nginx release, use the --from-version and --to-version flags: $ kubectl ingress-nginx lint --all-namespaces --verbose --from-version 0 .24.0 --to-version 0 .24.0 Checking ingresses... \u2717 anamespace/this-nginx - Contains the removed session-cookie-hash annotation. Lint added for version 0.24.0 https://github.com/kubernetes/ingress-nginx/issues/3743 Checking deployments... \u2717 namespace2/nginx-ingress-controller - Uses removed config flag --enable-dynamic-certificates Lint added for version 0.24.0 https://github.com/kubernetes/ingress-nginx/issues/3808 logs \u00b6 kubectl ingress-nginx logs is almost the same as kubectl logs , with fewer flags. It will automatically choose an ingress-nginx pod to read logs from. $ kubectl ingress-nginx logs -n ingress-nginx ------------------------------------------------------------------------------- NGINX Ingress controller Release: dev Build: git-48dc3a867 Repository: git@github.com:kubernetes/ingress-nginx.git ------------------------------------------------------------------------------- W0405 16:53:46.061589 7 flags.go:214] SSL certificate chain completion is disabled (--enable-ssl-chain-completion=false) nginx version: nginx/1.15.9 W0405 16:53:46.070093 7 client_config.go:549] Neither --kubeconfig nor --master was specified. Using the inClusterConfig. This might not work. I0405 16:53:46.070499 7 main.go:205] Creating API client for https://10.96.0.1:443 I0405 16:53:46.077784 7 main.go:249] Running in Kubernetes cluster version v1.10 (v1.10.11) - git (clean) commit 637c7e288581ee40ab4ca210618a89a555b6e7e9 - platform linux/amd64 I0405 16:53:46.183359 7 nginx.go:265] Starting NGINX Ingress controller I0405 16:53:46.193913 7 event.go:209] Event(v1.ObjectReference{Kind:\"ConfigMap\", Namespace:\"ingress-nginx\", Name:\"udp-services\", UID:\"82258915-563e-11e9-9c52-025000000001\", APIVersion:\"v1\", ResourceVersion:\"494\", FieldPath:\"\"}): type: 'Normal' reason: 'CREATE' ConfigMap ingress-nginx/udp-services ... ssh \u00b6 kubectl ingress-nginx ssh is exactly the same as kubectl ingress-nginx exec -it -- /bin/bash . Use it when you want to quickly be dropped into a shell inside a running ingress-nginx container. $ kubectl ingress-nginx ssh -n ingress-nginx www-data@nginx-ingress-controller-7cbf77c976-wx5pn:/etc/nginx$","title":"kubectl plugin"},{"location":"kubectl-plugin/#the-ingress-nginx-kubectl-plugin","text":"","title":"The ingress-nginx kubectl plugin"},{"location":"kubectl-plugin/#installation","text":"Install krew , then run kubectl krew install ingress-nginx to install the plugin. Then run kubectl ingress-nginx --help to make sure the plugin is properly installed and to get a list of commands: kubectl ingress-nginx --help A kubectl plugin for inspecting your ingress-nginx deployments Usage: ingress-nginx [command] Available Commands: backends Inspect the dynamic backend information of an ingress-nginx instance certs Output the certificate data stored in an ingress-nginx pod conf Inspect the generated nginx.conf exec Execute a command inside an ingress-nginx pod general Inspect the other dynamic ingress-nginx information help Help about any command info Show information about the ingress-nginx service ingresses Provide a short summary of all of the ingress definitions lint Inspect kubernetes resources for possible issues logs Get the kubernetes logs for an ingress-nginx pod ssh ssh into a running ingress-nginx pod Flags: --as string Username to impersonate for the operation --as-group stringArray Group to impersonate for the operation, this flag can be repeated to specify multiple groups. --cache-dir string Default HTTP cache directory (default \"/Users/alexkursell/.kube/http-cache\") --certificate-authority string Path to a cert file for the certificate authority --client-certificate string Path to a client certificate file for TLS --client-key string Path to a client key file for TLS --cluster string The name of the kubeconfig cluster to use --context string The name of the kubeconfig context to use -h, --help help for ingress-nginx --insecure-skip-tls-verify If true, the server's certificate will not be checked for validity. This will make your HTTPS connections insecure --kubeconfig string Path to the kubeconfig file to use for CLI requests. -n, --namespace string If present, the namespace scope for this CLI request --request-timeout string The length of time to wait before giving up on a single server request. Non-zero values should contain a corresponding time unit (e.g. 1s, 2m, 3h). A value of zero means don't timeout requests. (default \"0\") -s, --server string The address and port of the Kubernetes API server --token string Bearer token for authentication to the API server --user string The name of the kubeconfig user to use Use \"ingress-nginx [command] --help\" for more information about a command.","title":"Installation"},{"location":"kubectl-plugin/#common-flags","text":"Every subcommand supports the basic kubectl configuration flags like --namespace , --context , --client-key and so on. Subcommands that act on a particular ingress-nginx pod ( backends , certs , conf , exec , general , logs , ssh ), support the --deployment and --pod flags to select either a pod from a deployment with the given name, or a pod with the given name. The --deployment flag defaults to nginx-ingress-controller . Subcommands that inspect resources ( ingresses , lint ) support the --all-namespaces flag, which causes them to inspect resources in every namespace.","title":"Common Flags"},{"location":"kubectl-plugin/#subcommands","text":"Note that backends , general , certs , and conf require ingress-nginx version 0.23.0 or higher.","title":"Subcommands"},{"location":"kubectl-plugin/#backends","text":"Run kubectl ingress-nginx backends to get a JSON array of the backends that an ingress-nginx controller currently knows about: $ kubectl ingress-nginx backends -n ingress-nginx [ { \"name\": \"default-apple-service-5678\", \"service\": { \"metadata\": { \"creationTimestamp\": null }, \"spec\": { \"ports\": [ { \"protocol\": \"TCP\", \"port\": 5678, \"targetPort\": 5678 } ], \"selector\": { \"app\": \"apple\" }, \"clusterIP\": \"10.97.230.121\", \"type\": \"ClusterIP\", \"sessionAffinity\": \"None\" }, \"status\": { \"loadBalancer\": {} } }, \"port\": 0, \"sslPassthrough\": false, \"endpoints\": [ { \"address\": \"10.1.3.86\", \"port\": \"5678\" } ], \"sessionAffinityConfig\": { \"name\": \"\", \"cookieSessionAffinity\": { \"name\": \"\" } }, \"upstreamHashByConfig\": { \"upstream-hash-by-subset-size\": 3 }, \"noServer\": false, \"trafficShapingPolicy\": { \"weight\": 0, \"header\": \"\", \"headerValue\": \"\", \"cookie\": \"\" } }, { \"name\": \"default-echo-service-8080\", ... }, { \"name\": \"upstream-default-backend\", ... } ] Add the --list option to show only the backend names. Add the --backend option to show only the backend with the given name.","title":"backends"},{"location":"kubectl-plugin/#certs","text":"Use kubectl ingress-nginx certs --host to dump the SSL cert/key information for a given host. WARNING: This command will dump sensitive private key information. Don't blindly share the output, and certainly don't log it anywhere. $ kubectl ingress-nginx certs -n ingress-nginx --host testaddr.local -----BEGIN CERTIFICATE----- ... -----END CERTIFICATE----- -----BEGIN CERTIFICATE----- ... -----END CERTIFICATE----- -----BEGIN RSA PRIVATE KEY----- -----END RSA PRIVATE KEY-----","title":"certs"},{"location":"kubectl-plugin/#conf","text":"Use kubectl ingress-nginx conf to dump the generated nginx.conf file. Add the --host option to view only the server block for that host: kubectl ingress-nginx conf -n ingress-nginx --host testaddr.local server { server_name testaddr.local ; listen 80; set $proxy_upstream_name \"-\"; set $pass_access_scheme $scheme; set $pass_server_port $server_port; set $best_http_host $http_host; set $pass_port $pass_server_port; location / { set $namespace \"\"; set $ingress_name \"\"; set $service_name \"\"; set $service_port \"0\"; set $location_path \"/\"; ...","title":"conf"},{"location":"kubectl-plugin/#exec","text":"kubectl ingress-nginx exec is exactly the same as kubectl exec , with the same command flags. It will automatically choose an ingress-nginx pod to run the command in. $ kubectl ingress-nginx exec -i -n ingress-nginx -- ls /etc/nginx fastcgi_params geoip lua mime.types modsecurity modules nginx.conf opentracing.json owasp-modsecurity-crs template","title":"exec"},{"location":"kubectl-plugin/#general","text":"kubectl ingress-nginx general dumps miscellaneous controller state as a JSON object. Currently it just shows the number of controller pods known to a particular controller pod. $ kubectl ingress-nginx general -n ingress-nginx { \"controllerPodsCount\": 1 }","title":"general"},{"location":"kubectl-plugin/#info","text":"Shows the internal and external IP/CNAMES for an ingress-nginx service. $ kubectl ingress-nginx info -n ingress-nginx Service cluster IP address: 10.187.253.31 LoadBalancer IP|CNAME: 35.123.123.123 Use the --service flag if your ingress-nginx LoadBalancer service is not named ingress-nginx .","title":"info"},{"location":"kubectl-plugin/#ingresses","text":"kubectl ingress-nginx ingresses , alternately kubectl ingress-nginx ing , shows a more detailed view of the ingress definitions in a namespace. Compare: $ kubectl get ingresses --all-namespaces NAMESPACE NAME HOSTS ADDRESS PORTS AGE default example-ingress1 testaddr.local,testaddr2.local localhost 80 5d default test-ingress-2 * localhost 80 5d vs $ kubectl ingress-nginx ingresses --all-namespaces NAMESPACE INGRESS NAME HOST+PATH ADDRESSES TLS SERVICE SERVICE PORT ENDPOINTS default example-ingress1 testaddr.local/etameta localhost NO pear-service 5678 5 default example-ingress1 testaddr2.local/otherpath localhost NO apple-service 5678 1 default example-ingress1 testaddr2.local/otherotherpath localhost NO pear-service 5678 5 default test-ingress-2 * localhost NO echo-service 8080 2","title":"ingresses"},{"location":"kubectl-plugin/#lint","text":"kubectl ingress-nginx lint can check a namespace or entire cluster for potential configuration issues. This command is especially useful when upgrading between ingress-nginx versions. $ kubectl ingress-nginx lint --all-namespaces --verbose Checking ingresses... \u2717 anamespace/this-nginx - Contains the removed session-cookie-hash annotation. Lint added for version 0.24.0 https://github.com/kubernetes/ingress-nginx/issues/3743 \u2717 othernamespace/ingress-definition-blah - The rewrite-target annotation value does not reference a capture group Lint added for version 0.22.0 https://github.com/kubernetes/ingress-nginx/issues/3174 Checking deployments... \u2717 namespace2/nginx-ingress-controller - Uses removed config flag --sort-backends Lint added for version 0.22.0 https://github.com/kubernetes/ingress-nginx/issues/3655 - Uses removed config flag --enable-dynamic-certificates Lint added for version 0.24.0 https://github.com/kubernetes/ingress-nginx/issues/3808 to show the lints added only for a particular ingress-nginx release, use the --from-version and --to-version flags: $ kubectl ingress-nginx lint --all-namespaces --verbose --from-version 0 .24.0 --to-version 0 .24.0 Checking ingresses... \u2717 anamespace/this-nginx - Contains the removed session-cookie-hash annotation. Lint added for version 0.24.0 https://github.com/kubernetes/ingress-nginx/issues/3743 Checking deployments... \u2717 namespace2/nginx-ingress-controller - Uses removed config flag --enable-dynamic-certificates Lint added for version 0.24.0 https://github.com/kubernetes/ingress-nginx/issues/3808","title":"lint"},{"location":"kubectl-plugin/#logs","text":"kubectl ingress-nginx logs is almost the same as kubectl logs , with fewer flags. It will automatically choose an ingress-nginx pod to read logs from. $ kubectl ingress-nginx logs -n ingress-nginx ------------------------------------------------------------------------------- NGINX Ingress controller Release: dev Build: git-48dc3a867 Repository: git@github.com:kubernetes/ingress-nginx.git ------------------------------------------------------------------------------- W0405 16:53:46.061589 7 flags.go:214] SSL certificate chain completion is disabled (--enable-ssl-chain-completion=false) nginx version: nginx/1.15.9 W0405 16:53:46.070093 7 client_config.go:549] Neither --kubeconfig nor --master was specified. Using the inClusterConfig. This might not work. I0405 16:53:46.070499 7 main.go:205] Creating API client for https://10.96.0.1:443 I0405 16:53:46.077784 7 main.go:249] Running in Kubernetes cluster version v1.10 (v1.10.11) - git (clean) commit 637c7e288581ee40ab4ca210618a89a555b6e7e9 - platform linux/amd64 I0405 16:53:46.183359 7 nginx.go:265] Starting NGINX Ingress controller I0405 16:53:46.193913 7 event.go:209] Event(v1.ObjectReference{Kind:\"ConfigMap\", Namespace:\"ingress-nginx\", Name:\"udp-services\", UID:\"82258915-563e-11e9-9c52-025000000001\", APIVersion:\"v1\", ResourceVersion:\"494\", FieldPath:\"\"}): type: 'Normal' reason: 'CREATE' ConfigMap ingress-nginx/udp-services ...","title":"logs"},{"location":"kubectl-plugin/#ssh","text":"kubectl ingress-nginx ssh is exactly the same as kubectl ingress-nginx exec -it -- /bin/bash . Use it when you want to quickly be dropped into a shell inside a running ingress-nginx container. $ kubectl ingress-nginx ssh -n ingress-nginx www-data@nginx-ingress-controller-7cbf77c976-wx5pn:/etc/nginx$","title":"ssh"},{"location":"troubleshooting/","text":"Troubleshooting \u00b6 Ingress-Controller Logs and Events \u00b6 There are many ways to troubleshoot the ingress-controller. The following are basic troubleshooting methods to obtain more information. Check the Ingress Resource Events $ kubectl get ing -n NAME HOSTS ADDRESS PORTS AGE cafe-ingress cafe.com 10.0.2.15 80 25s $ kubectl describe ing -n Name: cafe-ingress Namespace: default Address: 10.0.2.15 Default backend: default-http-backend:80 (172.17.0.5:8080) Rules: Host Path Backends ---- ---- -------- cafe.com /tea tea-svc:80 () /coffee coffee-svc:80 () Annotations: kubectl.kubernetes.io/last-applied-configuration: {\"apiVersion\":\"networking.k8s.io/v1beta1\",\"kind\":\"Ingress\",\"metadata\":{\"annotations\":{},\"name\":\"cafe-ingress\",\"namespace\":\"default\",\"selfLink\":\"/apis/networking/v1beta1/namespaces/default/ingresses/cafe-ingress\"},\"spec\":{\"rules\":[{\"host\":\"cafe.com\",\"http\":{\"paths\":[{\"backend\":{\"serviceName\":\"tea-svc\",\"servicePort\":80},\"path\":\"/tea\"},{\"backend\":{\"serviceName\":\"coffee-svc\",\"servicePort\":80},\"path\":\"/coffee\"}]}}]},\"status\":{\"loadBalancer\":{\"ingress\":[{\"ip\":\"169.48.142.110\"}]}}} Events: Type Reason Age From Message ---- ------ ---- ---- ------- Normal CREATE 1m nginx-ingress-controller Ingress default/cafe-ingress Normal UPDATE 58s nginx-ingress-controller Ingress default/cafe-ingress Check the Ingress Controller Logs $ kubectl get pods -n NAME READY STATUS RESTARTS AGE nginx-ingress-controller-67956bf89d-fv58j 1/1 Running 0 1m $ kubectl logs -n nginx-ingress-controller-67956bf89d-fv58j ------------------------------------------------------------------------------- NGINX Ingress controller Release: 0.14.0 Build: git-734361d Repository: https://github.com/kubernetes/ingress-nginx ------------------------------------------------------------------------------- .... Check the Nginx Configuration $ kubectl get pods -n NAME READY STATUS RESTARTS AGE nginx-ingress-controller-67956bf89d-fv58j 1/1 Running 0 1m $ kubectl exec -it -n nginx-ingress-controller-67956bf89d-fv58j -- cat /etc/nginx/nginx.conf daemon off; worker_processes 2; pid /run/nginx.pid; worker_rlimit_nofile 523264; worker_shutdown_timeout 240s; events { multi_accept on; worker_connections 16384; use epoll; } http { .... Check if used Services Exist $ kubectl get svc --all-namespaces NAMESPACE NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE default coffee-svc ClusterIP 10.106.154.35 80/TCP 18m default kubernetes ClusterIP 10.96.0.1 443/TCP 30m default tea-svc ClusterIP 10.104.172.12 80/TCP 18m kube-system default-http-backend NodePort 10.108.189.236 80:30001/TCP 30m kube-system kube-dns ClusterIP 10.96.0.10 53/UDP,53/TCP 30m kube-system kubernetes-dashboard NodePort 10.103.128.17 80:30000/TCP 30m Debug Logging \u00b6 Using the flag --v=XX it is possible to increase the level of logging. This is performed by editing the deployment. $ kubectl get deploy -n NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE default-http-backend 1 1 1 1 35m nginx-ingress-controller 1 1 1 1 35m $ kubectl edit deploy -n nginx-ingress-controller # Add --v = X to \"- args\" , where X is an integer --v=2 shows details using diff about the changes in the configuration in nginx --v=3 shows details about the service, Ingress rule, endpoint changes and it dumps the nginx configuration in JSON format --v=5 configures NGINX in debug mode Authentication to the Kubernetes API Server \u00b6 A number of components are involved in the authentication process and the first step is to narrow down the source of the problem, namely whether it is a problem with service authentication or with the kubeconfig file. Both authentications must work: +-------------+ service +------------+ | | authentication | | + apiserver +<-------------------+ ingress | | | | controller | +-------------+ +------------+ Service authentication The Ingress controller needs information from apiserver. Therefore, authentication is required, which can be achieved in two different ways: Service Account: This is recommended, because nothing has to be configured. The Ingress controller will use information provided by the system to communicate with the API server. See 'Service Account' section for details. Kubeconfig file: In some Kubernetes environments service accounts are not available. In this case a manual configuration is required. The Ingress controller binary can be started with the --kubeconfig flag. The value of the flag is a path to a file specifying how to connect to the API server. Using the --kubeconfig does not requires the flag --apiserver-host . The format of the file is identical to ~/.kube/config which is used by kubectl to connect to the API server. See 'kubeconfig' section for details. Using the flag --apiserver-host : Using this flag --apiserver-host=http://localhost:8080 it is possible to specify an unsecured API server or reach a remote kubernetes cluster using kubectl proxy . Please do not use this approach in production. In the diagram below you can see the full authentication flow with all options, starting with the browser on the lower left hand side. Kubernetes Workstation +---------------------------------------------------+ +------------------+ | | | | | +-----------+ apiserver +------------+ | | +------------+ | | | | proxy | | | | | | | | | apiserver | | ingress | | | | ingress | | | | | | controller | | | | controller | | | | | | | | | | | | | | | | | | | | | | | | | service account/ | | | | | | | | | | kubeconfig | | | | | | | | | +<-------------------+ | | | | | | | | | | | | | | | | | +------+----+ kubeconfig +------+-----+ | | +------+-----+ | | |<--------------------------------------------------------| | | | | | +---------------------------------------------------+ +------------------+ Service Account \u00b6 If using a service account to connect to the API server, the ingress-controller expects the file /var/run/secrets/kubernetes.io/serviceaccount/token to be present. It provides a secret token that is required to authenticate with the API server. Verify with the following commands: # start a container that contains curl $ kubectl run test --image = tutum/curl -- sleep 10000 # check that container is running $ kubectl get pods NAME READY STATUS RESTARTS AGE test-701078429-s5kca 1/1 Running 0 16s # check if secret exists $ kubectl exec test-701078429-s5kca -- ls /var/run/secrets/kubernetes.io/serviceaccount/ ca.crt namespace token # get service IP of master $ kubectl get services NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE kubernetes 10.0.0.1 443/TCP 1d # check base connectivity from cluster inside $ kubectl exec test-701078429-s5kca -- curl -k https://10.0.0.1 Unauthorized # connect using tokens $ TOKEN_VALUE = $( kubectl exec test-701078429-s5kca -- cat /var/run/secrets/kubernetes.io/serviceaccount/token ) $ echo $TOKEN_VALUE eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3Mi....9A $ kubectl exec test-701078429-s5kca -- curl --cacert /var/run/secrets/kubernetes.io/serviceaccount/ca.crt -H \"Authorization: Bearer $TOKEN_VALUE \" https://10.0.0.1 { \"paths\": [ \"/api\", \"/api/v1\", \"/apis\", \"/apis/apps\", \"/apis/apps/v1alpha1\", \"/apis/authentication.k8s.io\", \"/apis/authentication.k8s.io/v1beta1\", \"/apis/authorization.k8s.io\", \"/apis/authorization.k8s.io/v1beta1\", \"/apis/autoscaling\", \"/apis/autoscaling/v1\", \"/apis/batch\", \"/apis/batch/v1\", \"/apis/batch/v2alpha1\", \"/apis/certificates.k8s.io\", \"/apis/certificates.k8s.io/v1alpha1\", \"/apis/networking\", \"/apis/networking/v1beta1\", \"/apis/policy\", \"/apis/policy/v1alpha1\", \"/apis/rbac.authorization.k8s.io\", \"/apis/rbac.authorization.k8s.io/v1alpha1\", \"/apis/storage.k8s.io\", \"/apis/storage.k8s.io/v1beta1\", \"/healthz\", \"/healthz/ping\", \"/logs\", \"/metrics\", \"/swaggerapi/\", \"/ui/\", \"/version\" ] } If it is not working, there are two possible reasons: The contents of the tokens are invalid. Find the secret name with kubectl get secrets | grep service-account and delete it with kubectl delete secret . It will automatically be recreated. You have a non-standard Kubernetes installation and the file containing the token may not be present. The API server will mount a volume containing this file, but only if the API server is configured to use the ServiceAccount admission controller. If you experience this error, verify that your API server is using the ServiceAccount admission controller. If you are configuring the API server by hand, you can set this with the --admission-control parameter. Note that you should use other admission controllers as well. Before configuring this option, you should read about admission controllers. More information: User Guide: Service Accounts Cluster Administrator Guide: Managing Service Accounts Kube-Config \u00b6 If you want to use a kubeconfig file for authentication, follow the deploy procedure and add the flag --kubeconfig=/etc/kubernetes/kubeconfig.yaml to the args section of the deployment. Using GDB with Nginx \u00b6 Gdb can be used to with nginx to perform a configuration dump. This allows us to see which configuration is being used, as well as older configurations. Note: The below is based on the nginx documentation . SSH into the worker $ ssh user@workerIP Obtain the Docker Container Running nginx $ docker ps | grep nginx-ingress-controller CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES d9e1d243156a quay.io/kubernetes-ingress-controller/nginx-ingress-controller \"/usr/bin/dumb-init \u2026\" 19 minutes ago Up 19 minutes k8s_nginx-ingress-controller_nginx-ingress-controller-67956bf89d-mqxzt_kube-system_079f31ec-aa37-11e8-ad39-080027a227db_0 Exec into the container $ docker exec -it --user = 0 --privileged d9e1d243156a bash Make sure nginx is running in --with-debug $ nginx -V 2 > & 1 | grep -- '--with-debug' Get list of processes running on container $ ps -ef UID PID PPID C STIME TTY TIME CMD root 1 0 0 20:23 ? 00:00:00 /usr/bin/dumb-init /nginx-ingres root 5 1 0 20:23 ? 00:00:05 /nginx-ingress-controller --defa root 21 5 0 20:23 ? 00:00:00 nginx: master process /usr/sbin/ nobody 106 21 0 20:23 ? 00:00:00 nginx: worker process nobody 107 21 0 20:23 ? 00:00:00 nginx: worker process root 172 0 0 20:43 pts/0 00:00:00 bash Attach gdb to the nginx master process $ gdb -p 21 .... Attaching to process 21 Reading symbols from /usr/sbin/nginx...done. .... (gdb) Copy and paste the following: set $cd = ngx_cycle->config_dump set $nelts = $cd.nelts set $elts = (ngx_conf_dump_t*)($cd.elts) while ($nelts-- > 0) set $name = $elts[$nelts]->name.data printf \"Dumping %s to nginx_conf.txt\\n\", $name append memory nginx_conf.txt \\ $ elts [ $nelts ] ->buffer.start $elts [ $nelts ] ->buffer.end end Quit GDB by pressing CTRL+D Open nginx_conf.txt cat nginx_conf.txt","title":"Troubleshooting"},{"location":"troubleshooting/#troubleshooting","text":"","title":"Troubleshooting"},{"location":"troubleshooting/#ingress-controller-logs-and-events","text":"There are many ways to troubleshoot the ingress-controller. The following are basic troubleshooting methods to obtain more information. Check the Ingress Resource Events $ kubectl get ing -n NAME HOSTS ADDRESS PORTS AGE cafe-ingress cafe.com 10.0.2.15 80 25s $ kubectl describe ing -n Name: cafe-ingress Namespace: default Address: 10.0.2.15 Default backend: default-http-backend:80 (172.17.0.5:8080) Rules: Host Path Backends ---- ---- -------- cafe.com /tea tea-svc:80 () /coffee coffee-svc:80 () Annotations: kubectl.kubernetes.io/last-applied-configuration: {\"apiVersion\":\"networking.k8s.io/v1beta1\",\"kind\":\"Ingress\",\"metadata\":{\"annotations\":{},\"name\":\"cafe-ingress\",\"namespace\":\"default\",\"selfLink\":\"/apis/networking/v1beta1/namespaces/default/ingresses/cafe-ingress\"},\"spec\":{\"rules\":[{\"host\":\"cafe.com\",\"http\":{\"paths\":[{\"backend\":{\"serviceName\":\"tea-svc\",\"servicePort\":80},\"path\":\"/tea\"},{\"backend\":{\"serviceName\":\"coffee-svc\",\"servicePort\":80},\"path\":\"/coffee\"}]}}]},\"status\":{\"loadBalancer\":{\"ingress\":[{\"ip\":\"169.48.142.110\"}]}}} Events: Type Reason Age From Message ---- ------ ---- ---- ------- Normal CREATE 1m nginx-ingress-controller Ingress default/cafe-ingress Normal UPDATE 58s nginx-ingress-controller Ingress default/cafe-ingress Check the Ingress Controller Logs $ kubectl get pods -n NAME READY STATUS RESTARTS AGE nginx-ingress-controller-67956bf89d-fv58j 1/1 Running 0 1m $ kubectl logs -n nginx-ingress-controller-67956bf89d-fv58j ------------------------------------------------------------------------------- NGINX Ingress controller Release: 0.14.0 Build: git-734361d Repository: https://github.com/kubernetes/ingress-nginx ------------------------------------------------------------------------------- .... Check the Nginx Configuration $ kubectl get pods -n NAME READY STATUS RESTARTS AGE nginx-ingress-controller-67956bf89d-fv58j 1/1 Running 0 1m $ kubectl exec -it -n nginx-ingress-controller-67956bf89d-fv58j -- cat /etc/nginx/nginx.conf daemon off; worker_processes 2; pid /run/nginx.pid; worker_rlimit_nofile 523264; worker_shutdown_timeout 240s; events { multi_accept on; worker_connections 16384; use epoll; } http { .... Check if used Services Exist $ kubectl get svc --all-namespaces NAMESPACE NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE default coffee-svc ClusterIP 10.106.154.35 80/TCP 18m default kubernetes ClusterIP 10.96.0.1 443/TCP 30m default tea-svc ClusterIP 10.104.172.12 80/TCP 18m kube-system default-http-backend NodePort 10.108.189.236 80:30001/TCP 30m kube-system kube-dns ClusterIP 10.96.0.10 53/UDP,53/TCP 30m kube-system kubernetes-dashboard NodePort 10.103.128.17 80:30000/TCP 30m","title":"Ingress-Controller Logs and Events"},{"location":"troubleshooting/#debug-logging","text":"Using the flag --v=XX it is possible to increase the level of logging. This is performed by editing the deployment. $ kubectl get deploy -n NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE default-http-backend 1 1 1 1 35m nginx-ingress-controller 1 1 1 1 35m $ kubectl edit deploy -n nginx-ingress-controller # Add --v = X to \"- args\" , where X is an integer --v=2 shows details using diff about the changes in the configuration in nginx --v=3 shows details about the service, Ingress rule, endpoint changes and it dumps the nginx configuration in JSON format --v=5 configures NGINX in debug mode","title":"Debug Logging"},{"location":"troubleshooting/#authentication-to-the-kubernetes-api-server","text":"A number of components are involved in the authentication process and the first step is to narrow down the source of the problem, namely whether it is a problem with service authentication or with the kubeconfig file. Both authentications must work: +-------------+ service +------------+ | | authentication | | + apiserver +<-------------------+ ingress | | | | controller | +-------------+ +------------+ Service authentication The Ingress controller needs information from apiserver. Therefore, authentication is required, which can be achieved in two different ways: Service Account: This is recommended, because nothing has to be configured. The Ingress controller will use information provided by the system to communicate with the API server. See 'Service Account' section for details. Kubeconfig file: In some Kubernetes environments service accounts are not available. In this case a manual configuration is required. The Ingress controller binary can be started with the --kubeconfig flag. The value of the flag is a path to a file specifying how to connect to the API server. Using the --kubeconfig does not requires the flag --apiserver-host . The format of the file is identical to ~/.kube/config which is used by kubectl to connect to the API server. See 'kubeconfig' section for details. Using the flag --apiserver-host : Using this flag --apiserver-host=http://localhost:8080 it is possible to specify an unsecured API server or reach a remote kubernetes cluster using kubectl proxy . Please do not use this approach in production. In the diagram below you can see the full authentication flow with all options, starting with the browser on the lower left hand side. Kubernetes Workstation +---------------------------------------------------+ +------------------+ | | | | | +-----------+ apiserver +------------+ | | +------------+ | | | | proxy | | | | | | | | | apiserver | | ingress | | | | ingress | | | | | | controller | | | | controller | | | | | | | | | | | | | | | | | | | | | | | | | service account/ | | | | | | | | | | kubeconfig | | | | | | | | | +<-------------------+ | | | | | | | | | | | | | | | | | +------+----+ kubeconfig +------+-----+ | | +------+-----+ | | |<--------------------------------------------------------| | | | | | +---------------------------------------------------+ +------------------+","title":"Authentication to the Kubernetes API Server"},{"location":"troubleshooting/#service-account","text":"If using a service account to connect to the API server, the ingress-controller expects the file /var/run/secrets/kubernetes.io/serviceaccount/token to be present. It provides a secret token that is required to authenticate with the API server. Verify with the following commands: # start a container that contains curl $ kubectl run test --image = tutum/curl -- sleep 10000 # check that container is running $ kubectl get pods NAME READY STATUS RESTARTS AGE test-701078429-s5kca 1/1 Running 0 16s # check if secret exists $ kubectl exec test-701078429-s5kca -- ls /var/run/secrets/kubernetes.io/serviceaccount/ ca.crt namespace token # get service IP of master $ kubectl get services NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE kubernetes 10.0.0.1 443/TCP 1d # check base connectivity from cluster inside $ kubectl exec test-701078429-s5kca -- curl -k https://10.0.0.1 Unauthorized # connect using tokens $ TOKEN_VALUE = $( kubectl exec test-701078429-s5kca -- cat /var/run/secrets/kubernetes.io/serviceaccount/token ) $ echo $TOKEN_VALUE eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3Mi....9A $ kubectl exec test-701078429-s5kca -- curl --cacert /var/run/secrets/kubernetes.io/serviceaccount/ca.crt -H \"Authorization: Bearer $TOKEN_VALUE \" https://10.0.0.1 { \"paths\": [ \"/api\", \"/api/v1\", \"/apis\", \"/apis/apps\", \"/apis/apps/v1alpha1\", \"/apis/authentication.k8s.io\", \"/apis/authentication.k8s.io/v1beta1\", \"/apis/authorization.k8s.io\", \"/apis/authorization.k8s.io/v1beta1\", \"/apis/autoscaling\", \"/apis/autoscaling/v1\", \"/apis/batch\", \"/apis/batch/v1\", \"/apis/batch/v2alpha1\", \"/apis/certificates.k8s.io\", \"/apis/certificates.k8s.io/v1alpha1\", \"/apis/networking\", \"/apis/networking/v1beta1\", \"/apis/policy\", \"/apis/policy/v1alpha1\", \"/apis/rbac.authorization.k8s.io\", \"/apis/rbac.authorization.k8s.io/v1alpha1\", \"/apis/storage.k8s.io\", \"/apis/storage.k8s.io/v1beta1\", \"/healthz\", \"/healthz/ping\", \"/logs\", \"/metrics\", \"/swaggerapi/\", \"/ui/\", \"/version\" ] } If it is not working, there are two possible reasons: The contents of the tokens are invalid. Find the secret name with kubectl get secrets | grep service-account and delete it with kubectl delete secret . It will automatically be recreated. You have a non-standard Kubernetes installation and the file containing the token may not be present. The API server will mount a volume containing this file, but only if the API server is configured to use the ServiceAccount admission controller. If you experience this error, verify that your API server is using the ServiceAccount admission controller. If you are configuring the API server by hand, you can set this with the --admission-control parameter. Note that you should use other admission controllers as well. Before configuring this option, you should read about admission controllers. More information: User Guide: Service Accounts Cluster Administrator Guide: Managing Service Accounts","title":"Service Account"},{"location":"troubleshooting/#kube-config","text":"If you want to use a kubeconfig file for authentication, follow the deploy procedure and add the flag --kubeconfig=/etc/kubernetes/kubeconfig.yaml to the args section of the deployment.","title":"Kube-Config"},{"location":"troubleshooting/#using-gdb-with-nginx","text":"Gdb can be used to with nginx to perform a configuration dump. This allows us to see which configuration is being used, as well as older configurations. Note: The below is based on the nginx documentation . SSH into the worker $ ssh user@workerIP Obtain the Docker Container Running nginx $ docker ps | grep nginx-ingress-controller CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES d9e1d243156a quay.io/kubernetes-ingress-controller/nginx-ingress-controller \"/usr/bin/dumb-init \u2026\" 19 minutes ago Up 19 minutes k8s_nginx-ingress-controller_nginx-ingress-controller-67956bf89d-mqxzt_kube-system_079f31ec-aa37-11e8-ad39-080027a227db_0 Exec into the container $ docker exec -it --user = 0 --privileged d9e1d243156a bash Make sure nginx is running in --with-debug $ nginx -V 2 > & 1 | grep -- '--with-debug' Get list of processes running on container $ ps -ef UID PID PPID C STIME TTY TIME CMD root 1 0 0 20:23 ? 00:00:00 /usr/bin/dumb-init /nginx-ingres root 5 1 0 20:23 ? 00:00:05 /nginx-ingress-controller --defa root 21 5 0 20:23 ? 00:00:00 nginx: master process /usr/sbin/ nobody 106 21 0 20:23 ? 00:00:00 nginx: worker process nobody 107 21 0 20:23 ? 00:00:00 nginx: worker process root 172 0 0 20:43 pts/0 00:00:00 bash Attach gdb to the nginx master process $ gdb -p 21 .... Attaching to process 21 Reading symbols from /usr/sbin/nginx...done. .... (gdb) Copy and paste the following: set $cd = ngx_cycle->config_dump set $nelts = $cd.nelts set $elts = (ngx_conf_dump_t*)($cd.elts) while ($nelts-- > 0) set $name = $elts[$nelts]->name.data printf \"Dumping %s to nginx_conf.txt\\n\", $name append memory nginx_conf.txt \\ $ elts [ $nelts ] ->buffer.start $elts [ $nelts ] ->buffer.end end Quit GDB by pressing CTRL+D Open nginx_conf.txt cat nginx_conf.txt","title":"Using GDB with Nginx"},{"location":"deploy/","text":"Installation Guide \u00b6 Attention The default configuration watches Ingress object from all the namespaces . To change this behavior use the flag --watch-namespace to limit the scope to a particular namespace. Warning If multiple Ingresses define paths for the same host, the ingress controller merges the definitions . Danger The admission webhook requires connectivity between Kubernetes API server and the ingress controller. In case Network policies or additional firewalls, please allow access to port 8443 . Attention The first time the ingress controller starts, two Jobs create the SSL Certificate used by the admission webhook. For this reason, there is an initial delay of up to two minutes until it is possible to create and validate Ingress definitions. You can wait until it is ready to run the next command: kubectl wait --namespace ingress-nginx \\ --for=condition=ready pod \\ --selector=app.kubernetes.io/component=controller \\ --timeout=120s Contents \u00b6 Provider Specific Steps Docker for Mac minikube AWS GCE - GKE Azure Digital Ocean Scaleway Bare-metal Verify installation Detect installed version Using Helm Provider Specific Steps \u00b6 Docker for Mac \u00b6 Kubernetes is available in Docker for Mac (from version 18.06.0-ce ) kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/cloud/deploy.yaml minikube \u00b6 For standard usage: minikube addons enable ingress AWS \u00b6 In AWS we use a Network load balancer (NLB) to expose the NGINX Ingress controller behind a Service of Type=LoadBalancer . Network Load Balancer (NLB) \u00b6 kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/aws/deploy.yaml TLS termination in AWS Load Balancer (ELB) \u00b6 In some scenarios is required to terminate TLS in the Load Balancer and not in the ingress controller. For this purpose we provide a template: Download deploy-tls-termination.yaml wget https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/aws/deploy-tls-termination.yaml Edit the file and change: VPC CIDR in use for the Kubernetes cluster: proxy-real-ip-cidr: XXX.XXX.XXX/XX AWS Certificate Manager (ACM) ID arn:aws:acm:us-west-2:XXXXXXXX:certificate/XXXXXX-XXXXXXX-XXXXXXX-XXXXXXXX Deploy the manifest: kubectl apply -f deploy-tls-termination.yaml NLB Idle Timeouts \u00b6 Idle timeout value for TCP flows is 350 seconds and cannot be modified . For this reason, you need to ensure the keepalive_timeout value is configured less than 350 seconds to work as expected. By default NGINX keepalive_timeout is set to 75s . More information with regards to timeouts can be found in the official AWS documentation GCE-GKE \u00b6 Info Initialize your user as a cluster-admin with the following command: kubectl create clusterrolebinding cluster-admin-binding \\ --clusterrole cluster-admin \\ --user $(gcloud config get-value account) Danger For private clusters, you will need to either add an additional firewall rule that allows master nodes access to port 8443/tcp on worker nodes, or change the existing rule that allows access to ports 80/tcp , 443/tcp and 10254/tcp to also allow access to port 8443/tcp . See the GKE documentation on adding rules and the Kubernetes issue for more detail. kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/cloud/deploy.yaml Failure Proxy protocol is not supported in GCE/GKE Azure \u00b6 kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/cloud/deploy.yaml Digital Ocean \u00b6 kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/do/deploy.yaml Scaleway \u00b6 kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/static/provider/scw/deploy.yaml Bare-metal \u00b6 Using NodePort : kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/baremetal/deploy.yaml Tip For extended notes regarding deployments on bare-metal, see Bare-metal considerations . Verify installation \u00b6 Info In minikube the ingress addon is installed in the namespace kube-system instead of ingress-nginx To check if the ingress controller pods have started, run the following command: kubectl get pods -n ingress-nginx \\ -l app.kubernetes.io/name=ingress-nginx --watch Once the ingress controller pods are running, you can cancel the command typing Ctrl+C . Now, you are ready to create your first ingress. Detect installed version \u00b6 To detect which version of the ingress controller is running, exec into the pod and run nginx-ingress-controller version command. POD_NAMESPACE=ingress-nginx POD_NAME=$(kubectl get pods -n $POD_NAMESPACE -l app.kubernetes.io/name=ingress-nginx --field-selector=status.phase=Running -o jsonpath='{.items[0].metadata.name}') kubectl exec -it $POD_NAME -n $POD_NAMESPACE -- /nginx-ingress-controller --version Using Helm \u00b6 NGINX Ingress controller can be installed via Helm using the chart from the project repository. To install the chart with the release name ingress-nginx : helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx helm install my-release ingress-nginx/ingress-nginx If you are using Helm 2 then specify release name using --name flag helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx/ helm install --name ingress-nginx ingress-nginx/ingress-nginx Detect installed version: \u00b6 POD_NAME=$(kubectl get pods -l app.kubernetes.io/name=ingress-nginx -o jsonpath='{.items[0].metadata.name}') kubectl exec -it $POD_NAME -- /nginx-ingress-controller --version","title":"Installation Guide"},{"location":"deploy/#installation-guide","text":"Attention The default configuration watches Ingress object from all the namespaces . To change this behavior use the flag --watch-namespace to limit the scope to a particular namespace. Warning If multiple Ingresses define paths for the same host, the ingress controller merges the definitions . Danger The admission webhook requires connectivity between Kubernetes API server and the ingress controller. In case Network policies or additional firewalls, please allow access to port 8443 . Attention The first time the ingress controller starts, two Jobs create the SSL Certificate used by the admission webhook. For this reason, there is an initial delay of up to two minutes until it is possible to create and validate Ingress definitions. You can wait until it is ready to run the next command: kubectl wait --namespace ingress-nginx \\ --for=condition=ready pod \\ --selector=app.kubernetes.io/component=controller \\ --timeout=120s","title":"Installation Guide"},{"location":"deploy/#contents","text":"Provider Specific Steps Docker for Mac minikube AWS GCE - GKE Azure Digital Ocean Scaleway Bare-metal Verify installation Detect installed version Using Helm","title":"Contents"},{"location":"deploy/#provider-specific-steps","text":"","title":"Provider Specific Steps"},{"location":"deploy/#docker-for-mac","text":"Kubernetes is available in Docker for Mac (from version 18.06.0-ce ) kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/cloud/deploy.yaml","title":"Docker for Mac"},{"location":"deploy/#minikube","text":"For standard usage: minikube addons enable ingress","title":"minikube"},{"location":"deploy/#aws","text":"In AWS we use a Network load balancer (NLB) to expose the NGINX Ingress controller behind a Service of Type=LoadBalancer .","title":"AWS"},{"location":"deploy/#network-load-balancer-nlb","text":"kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/aws/deploy.yaml","title":"Network Load Balancer (NLB)"},{"location":"deploy/#tls-termination-in-aws-load-balancer-elb","text":"In some scenarios is required to terminate TLS in the Load Balancer and not in the ingress controller. For this purpose we provide a template: Download deploy-tls-termination.yaml wget https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/aws/deploy-tls-termination.yaml Edit the file and change: VPC CIDR in use for the Kubernetes cluster: proxy-real-ip-cidr: XXX.XXX.XXX/XX AWS Certificate Manager (ACM) ID arn:aws:acm:us-west-2:XXXXXXXX:certificate/XXXXXX-XXXXXXX-XXXXXXX-XXXXXXXX Deploy the manifest: kubectl apply -f deploy-tls-termination.yaml","title":"TLS termination in AWS Load Balancer (ELB)"},{"location":"deploy/#nlb-idle-timeouts","text":"Idle timeout value for TCP flows is 350 seconds and cannot be modified . For this reason, you need to ensure the keepalive_timeout value is configured less than 350 seconds to work as expected. By default NGINX keepalive_timeout is set to 75s . More information with regards to timeouts can be found in the official AWS documentation","title":"NLB Idle Timeouts"},{"location":"deploy/#gce-gke","text":"Info Initialize your user as a cluster-admin with the following command: kubectl create clusterrolebinding cluster-admin-binding \\ --clusterrole cluster-admin \\ --user $(gcloud config get-value account) Danger For private clusters, you will need to either add an additional firewall rule that allows master nodes access to port 8443/tcp on worker nodes, or change the existing rule that allows access to ports 80/tcp , 443/tcp and 10254/tcp to also allow access to port 8443/tcp . See the GKE documentation on adding rules and the Kubernetes issue for more detail. kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/cloud/deploy.yaml Failure Proxy protocol is not supported in GCE/GKE","title":"GCE-GKE"},{"location":"deploy/#azure","text":"kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/cloud/deploy.yaml","title":"Azure"},{"location":"deploy/#digital-ocean","text":"kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/do/deploy.yaml","title":"Digital Ocean"},{"location":"deploy/#scaleway","text":"kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/static/provider/scw/deploy.yaml","title":"Scaleway"},{"location":"deploy/#bare-metal","text":"Using NodePort : kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.35.0/deploy/static/provider/baremetal/deploy.yaml Tip For extended notes regarding deployments on bare-metal, see Bare-metal considerations .","title":"Bare-metal"},{"location":"deploy/#verify-installation","text":"Info In minikube the ingress addon is installed in the namespace kube-system instead of ingress-nginx To check if the ingress controller pods have started, run the following command: kubectl get pods -n ingress-nginx \\ -l app.kubernetes.io/name=ingress-nginx --watch Once the ingress controller pods are running, you can cancel the command typing Ctrl+C . Now, you are ready to create your first ingress.","title":"Verify installation"},{"location":"deploy/#detect-installed-version","text":"To detect which version of the ingress controller is running, exec into the pod and run nginx-ingress-controller version command. POD_NAMESPACE=ingress-nginx POD_NAME=$(kubectl get pods -n $POD_NAMESPACE -l app.kubernetes.io/name=ingress-nginx --field-selector=status.phase=Running -o jsonpath='{.items[0].metadata.name}') kubectl exec -it $POD_NAME -n $POD_NAMESPACE -- /nginx-ingress-controller --version","title":"Detect installed version"},{"location":"deploy/#using-helm","text":"NGINX Ingress controller can be installed via Helm using the chart from the project repository. To install the chart with the release name ingress-nginx : helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx helm install my-release ingress-nginx/ingress-nginx If you are using Helm 2 then specify release name using --name flag helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx/ helm install --name ingress-nginx ingress-nginx/ingress-nginx","title":"Using Helm"},{"location":"deploy/#detect-installed-version_1","text":"POD_NAME=$(kubectl get pods -l app.kubernetes.io/name=ingress-nginx -o jsonpath='{.items[0].metadata.name}') kubectl exec -it $POD_NAME -- /nginx-ingress-controller --version","title":"Detect installed version:"},{"location":"deploy/baremetal/","text":"Bare-metal considerations \u00b6 In traditional cloud environments, where network load balancers are available on-demand, a single Kubernetes manifest suffices to provide a single point of contact to the NGINX Ingress controller to external clients and, indirectly, to any application running inside the cluster. Bare-metal environments lack this commodity, requiring a slightly different setup to offer the same kind of access to external consumers. The rest of this document describes a few recommended approaches to deploying the NGINX Ingress controller inside a Kubernetes cluster running on bare-metal. A pure software solution: MetalLB \u00b6 MetalLB provides a network load-balancer implementation for Kubernetes clusters that do not run on a supported cloud provider, effectively allowing the usage of LoadBalancer Services within any cluster. This section demonstrates how to use the Layer 2 configuration mode of MetalLB together with the NGINX Ingress controller in a Kubernetes cluster that has publicly accessible nodes . In this mode, one node attracts all the traffic for the ingress-nginx Service IP. See Traffic policies for more details. Note The description of other supported configuration modes is off-scope for this document. Warning MetalLB is currently in beta . Read about the Project maturity and make sure you inform yourself by reading the official documentation thoroughly. MetalLB can be deployed either with a simple Kubernetes manifest or with Helm. The rest of this example assumes MetalLB was deployed following the Installation instructions. MetalLB requires a pool of IP addresses in order to be able to take ownership of the ingress-nginx Service. This pool can be defined in a ConfigMap named config located in the same namespace as the MetalLB controller. This pool of IPs must be dedicated to MetalLB's use, you can't reuse the Kubernetes node IPs or IPs handed out by a DHCP server. Example Given the following 3-node Kubernetes cluster (the external IP is added as an example, in most bare-metal environments this value is ) $ kubectl get node NAME STATUS ROLES EXTERNAL-IP host-1 Ready master 203.0.113.1 host-2 Ready node 203.0.113.2 host-3 Ready node 203.0.113.3 After creating the following ConfigMap, MetalLB takes ownership of one of the IP addresses in the pool and updates the loadBalancer IP field of the ingress-nginx Service accordingly. apiVersion : v1 kind : ConfigMap metadata : namespace : metallb-system name : config data : config : | address-pools: - name: default protocol: layer2 addresses: - 203.0.113.10-203.0.113.15 $ kubectl -n ingress-nginx get svc NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) default-http-backend ClusterIP 10.0.64.249 80/TCP ingress-nginx LoadBalancer 10.0.220.217 203.0.113.10 80:30100/TCP,443:30101/TCP As soon as MetalLB sets the external IP address of the ingress-nginx LoadBalancer Service, the corresponding entries are created in the iptables NAT table and the node with the selected IP address starts responding to HTTP requests on the ports configured in the LoadBalancer Service: $ curl -D- http://203.0.113.3 -H 'Host: myapp.example.com' HTTP/1.1 200 OK Server: nginx/1.15.2 Tip In order to preserve the source IP address in HTTP requests sent to NGINX, it is necessary to use the Local traffic policy. Traffic policies are described in more details in Traffic policies as well as in the next section. Over a NodePort Service \u00b6 Due to its simplicity, this is the setup a user will deploy by default when following the steps described in the installation guide . Info A Service of type NodePort exposes, via the kube-proxy component, the same unprivileged port (default: 30000-32767) on every Kubernetes node, masters included. For more information, see Services . In this configuration, the NGINX container remains isolated from the host network. As a result, it can safely bind to any port, including the standard HTTP ports 80 and 443. However, due to the container namespace isolation, a client located outside the cluster network (e.g. on the public internet) is not able to access Ingress hosts directly on ports 80 and 443. Instead, the external client must append the NodePort allocated to the ingress-nginx Service to HTTP requests. Example Given the NodePort 30100 allocated to the ingress-nginx Service $ kubectl -n ingress-nginx get svc NAME TYPE CLUSTER-IP PORT(S) default-http-backend ClusterIP 10.0.64.249 80/TCP ingress-nginx NodePort 10.0.220.217 80:30100/TCP,443:30101/TCP and a Kubernetes node with the public IP address 203.0.113.2 (the external IP is added as an example, in most bare-metal environments this value is ) $ kubectl get node NAME STATUS ROLES EXTERNAL-IP host-1 Ready master 203.0.113.1 host-2 Ready node 203.0.113.2 host-3 Ready node 203.0.113.3 a client would reach an Ingress with host: myapp.example.com at http://myapp.example.com:30100 , where the myapp.example.com subdomain resolves to the 203.0.113.2 IP address. Impact on the host system While it may sound tempting to reconfigure the NodePort range using the --service-node-port-range API server flag to include unprivileged ports and be able to expose ports 80 and 443, doing so may result in unexpected issues including (but not limited to) the use of ports otherwise reserved to system daemons and the necessity to grant kube-proxy privileges it may otherwise not require. This practice is therefore discouraged . See the other approaches proposed in this page for alternatives. This approach has a few other limitations one ought to be aware of: Source IP address Services of type NodePort perform source address translation by default. This means the source IP of a HTTP request is always the IP address of the Kubernetes node that received the request from the perspective of NGINX. The recommended way to preserve the source IP in a NodePort setup is to set the value of the externalTrafficPolicy field of the ingress-nginx Service spec to Local ( example ). Warning This setting effectively drops packets sent to Kubernetes nodes which are not running any instance of the NGINX Ingress controller. Consider assigning NGINX Pods to specific nodes in order to control on what nodes the NGINX Ingress controller should be scheduled or not scheduled. Example In a Kubernetes cluster composed of 3 nodes (the external IP is added as an example, in most bare-metal environments this value is ) $ kubectl get node NAME STATUS ROLES EXTERNAL-IP host-1 Ready master 203.0.113.1 host-2 Ready node 203.0.113.2 host-3 Ready node 203.0.113.3 with a ingress-nginx-controller Deployment composed of 2 replicas $ kubectl -n ingress-nginx get pod -o wide NAME READY STATUS IP NODE default-http-backend-7c5bc89cc9-p86md 1/1 Running 172.17.1.1 host-2 ingress-nginx-controller-cf9ff8c96-8vvf8 1/1 Running 172.17.0.3 host-3 ingress-nginx-controller-cf9ff8c96-pxsds 1/1 Running 172.17.1.4 host-2 Requests sent to host-2 and host-3 would be forwarded to NGINX and original client's IP would be preserved, while requests to host-1 would get dropped because there is no NGINX replica running on that node. Ingress status Because NodePort Services do not get a LoadBalancerIP assigned by definition, the NGINX Ingress controller does not update the status of Ingress objects it manages . $ kubectl get ingress NAME HOSTS ADDRESS PORTS test-ingress myapp.example.com 80 Despite the fact there is no load balancer providing a public IP address to the NGINX Ingress controller, it is possible to force the status update of all managed Ingress objects by setting the externalIPs field of the ingress-nginx Service. Warning There is more to setting externalIPs than just enabling the NGINX Ingress controller to update the status of Ingress objects. Please read about this option in the Services page of official Kubernetes documentation as well as the section about External IPs in this document for more information. Example Given the following 3-node Kubernetes cluster (the external IP is added as an example, in most bare-metal environments this value is ) $ kubectl get node NAME STATUS ROLES EXTERNAL-IP host-1 Ready master 203.0.113.1 host-2 Ready node 203.0.113.2 host-3 Ready node 203.0.113.3 one could edit the ingress-nginx Service and add the following field to the object spec spec : externalIPs : - 203.0.113.1 - 203.0.113.2 - 203.0.113.3 which would in turn be reflected on Ingress objects as follows: $ kubectl get ingress -o wide NAME HOSTS ADDRESS PORTS test-ingress myapp.example.com 203.0.113.1,203.0.113.2,203.0.113.3 80 Redirects As NGINX is not aware of the port translation operated by the NodePort Service , backend applications are responsible for generating redirect URLs that take into account the URL used by external clients, including the NodePort. Example Redirects generated by NGINX, for instance HTTP to HTTPS or domain to www.domain , are generated without NodePort: $ curl -D- http://myapp.example.com:30100 ` HTTP/1.1 308 Permanent Redirect Server: nginx/1.15.2 Location: https://myapp.example.com/ #-> missing NodePort in HTTPS redirect Via the host network \u00b6 In a setup where there is no external load balancer available but using NodePorts is not an option, one can configure ingress-nginx Pods to use the network of the host they run on instead of a dedicated network namespace. The benefit of this approach is that the NGINX Ingress controller can bind ports 80 and 443 directly to Kubernetes nodes' network interfaces, without the extra network translation imposed by NodePort Services. Note This approach does not leverage any Service object to expose the NGINX Ingress controller. If the ingress-nginx Service exists in the target cluster, it is recommended to delete it . This can be achieved by enabling the hostNetwork option in the Pods' spec. template : spec : hostNetwork : true Security considerations Enabling this option exposes every system daemon to the NGINX Ingress controller on any network interface, including the host's loopback. Please evaluate the impact this may have on the security of your system carefully. Example Consider this ingress-nginx-controller Deployment composed of 2 replicas, NGINX Pods inherit from the IP address of their host instead of an internal Pod IP. $ kubectl -n ingress-nginx get pod -o wide NAME READY STATUS IP NODE default-http-backend-7c5bc89cc9-p86md 1/1 Running 172.17.1.1 host-2 ingress-nginx-controller-5b4cf5fc6-7lg6c 1/1 Running 203.0.113.3 host-3 ingress-nginx-controller-5b4cf5fc6-lzrls 1/1 Running 203.0.113.2 host-2 One major limitation of this deployment approach is that only a single NGINX Ingress controller Pod may be scheduled on each cluster node, because binding the same port multiple times on the same network interface is technically impossible. Pods that are unschedulable due to such situation fail with the following event: $ kubectl -n ingress-nginx describe pod ... Events: Type Reason From Message ---- ------ ---- ------- Warning FailedScheduling default-scheduler 0/3 nodes are available: 3 node(s) didn't have free ports for the requested pod ports. One way to ensure only schedulable Pods are created is to deploy the NGINX Ingress controller as a DaemonSet instead of a traditional Deployment. Info A DaemonSet schedules exactly one type of Pod per cluster node, masters included, unless a node is configured to repel those Pods . For more information, see DaemonSet . Because most properties of DaemonSet objects are identical to Deployment objects, this documentation page leaves the configuration of the corresponding manifest at the user's discretion. Like with NodePorts, this approach has a few quirks it is important to be aware of. DNS resolution Pods configured with hostNetwork: true do not use the internal DNS resolver (i.e. kube-dns or CoreDNS ), unless their dnsPolicy spec field is set to ClusterFirstWithHostNet . Consider using this setting if NGINX is expected to resolve internal names for any reason. Ingress status Because there is no Service exposing the NGINX Ingress controller in a configuration using the host network, the default --publish-service flag used in standard cloud setups does not apply and the status of all Ingress objects remains blank. $ kubectl get ingress NAME HOSTS ADDRESS PORTS test-ingress myapp.example.com 80 Instead, and because bare-metal nodes usually don't have an ExternalIP, one has to enable the --report-node-internal-ip-address flag, which sets the status of all Ingress objects to the internal IP address of all nodes running the NGINX Ingress controller. Example Given a ingress-nginx-controller DaemonSet composed of 2 replicas $ kubectl -n ingress-nginx get pod -o wide NAME READY STATUS IP NODE default-http-backend-7c5bc89cc9-p86md 1/1 Running 172.17.1.1 host-2 ingress-nginx-controller-5b4cf5fc6-7lg6c 1/1 Running 203.0.113.3 host-3 ingress-nginx-controller-5b4cf5fc6-lzrls 1/1 Running 203.0.113.2 host-2 the controller sets the status of all Ingress objects it manages to the following value: $ kubectl get ingress -o wide NAME HOSTS ADDRESS PORTS test-ingress myapp.example.com 203.0.113.2,203.0.113.3 80 Note Alternatively, it is possible to override the address written to Ingress objects using the --publish-status-address flag. See Command line arguments . Using a self-provisioned edge \u00b6 Similarly to cloud environments, this deployment approach requires an edge network component providing a public entrypoint to the Kubernetes cluster. This edge component can be either hardware (e.g. vendor appliance) or software (e.g. HAproxy ) and is usually managed outside of the Kubernetes landscape by operations teams. Such deployment builds upon the NodePort Service described above in Over a NodePort Service , with one significant difference: external clients do not access cluster nodes directly, only the edge component does. This is particularly suitable for private Kubernetes clusters where none of the nodes has a public IP address. On the edge side, the only prerequisite is to dedicate a public IP address that forwards all HTTP traffic to Kubernetes nodes and/or masters. Incoming traffic on TCP ports 80 and 443 is forwarded to the corresponding HTTP and HTTPS NodePort on the target nodes as shown in the diagram below: External IPs \u00b6 Source IP address This method does not allow preserving the source IP of HTTP requests in any manner, it is therefore not recommended to use it despite its apparent simplicity. The externalIPs Service option was previously mentioned in the NodePort section. As per the Services page of the official Kubernetes documentation, the externalIPs option causes kube-proxy to route traffic sent to arbitrary IP addresses and on the Service ports to the endpoints of that Service. These IP addresses must belong to the target node . Example Given the following 3-node Kubernetes cluster (the external IP is added as an example, in most bare-metal environments this value is ) $ kubectl get node NAME STATUS ROLES EXTERNAL-IP host-1 Ready master 203.0.113.1 host-2 Ready node 203.0.113.2 host-3 Ready node 203.0.113.3 and the following ingress-nginx NodePort Service $ kubectl -n ingress-nginx get svc NAME TYPE CLUSTER-IP PORT(S) ingress-nginx NodePort 10.0.220.217 80:30100/TCP,443:30101/TCP One could set the following external IPs in the Service spec, and NGINX would become available on both the NodePort and the Service port: spec : externalIPs : - 203.0.113.2 - 203.0.113.3 $ curl -D- http://myapp.example.com:30100 HTTP/1.1 200 OK Server: nginx/1.15.2 $ curl -D- http://myapp.example.com HTTP/1.1 200 OK Server: nginx/1.15.2 We assume the myapp.example.com subdomain above resolves to both 203.0.113.2 and 203.0.113.3 IP addresses.","title":"Bare-metal considerations"},{"location":"deploy/baremetal/#bare-metal-considerations","text":"In traditional cloud environments, where network load balancers are available on-demand, a single Kubernetes manifest suffices to provide a single point of contact to the NGINX Ingress controller to external clients and, indirectly, to any application running inside the cluster. Bare-metal environments lack this commodity, requiring a slightly different setup to offer the same kind of access to external consumers. The rest of this document describes a few recommended approaches to deploying the NGINX Ingress controller inside a Kubernetes cluster running on bare-metal.","title":"Bare-metal considerations"},{"location":"deploy/baremetal/#a-pure-software-solution-metallb","text":"MetalLB provides a network load-balancer implementation for Kubernetes clusters that do not run on a supported cloud provider, effectively allowing the usage of LoadBalancer Services within any cluster. This section demonstrates how to use the Layer 2 configuration mode of MetalLB together with the NGINX Ingress controller in a Kubernetes cluster that has publicly accessible nodes . In this mode, one node attracts all the traffic for the ingress-nginx Service IP. See Traffic policies for more details. Note The description of other supported configuration modes is off-scope for this document. Warning MetalLB is currently in beta . Read about the Project maturity and make sure you inform yourself by reading the official documentation thoroughly. MetalLB can be deployed either with a simple Kubernetes manifest or with Helm. The rest of this example assumes MetalLB was deployed following the Installation instructions. MetalLB requires a pool of IP addresses in order to be able to take ownership of the ingress-nginx Service. This pool can be defined in a ConfigMap named config located in the same namespace as the MetalLB controller. This pool of IPs must be dedicated to MetalLB's use, you can't reuse the Kubernetes node IPs or IPs handed out by a DHCP server. Example Given the following 3-node Kubernetes cluster (the external IP is added as an example, in most bare-metal environments this value is ) $ kubectl get node NAME STATUS ROLES EXTERNAL-IP host-1 Ready master 203.0.113.1 host-2 Ready node 203.0.113.2 host-3 Ready node 203.0.113.3 After creating the following ConfigMap, MetalLB takes ownership of one of the IP addresses in the pool and updates the loadBalancer IP field of the ingress-nginx Service accordingly. apiVersion : v1 kind : ConfigMap metadata : namespace : metallb-system name : config data : config : | address-pools: - name: default protocol: layer2 addresses: - 203.0.113.10-203.0.113.15 $ kubectl -n ingress-nginx get svc NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) default-http-backend ClusterIP 10.0.64.249 80/TCP ingress-nginx LoadBalancer 10.0.220.217 203.0.113.10 80:30100/TCP,443:30101/TCP As soon as MetalLB sets the external IP address of the ingress-nginx LoadBalancer Service, the corresponding entries are created in the iptables NAT table and the node with the selected IP address starts responding to HTTP requests on the ports configured in the LoadBalancer Service: $ curl -D- http://203.0.113.3 -H 'Host: myapp.example.com' HTTP/1.1 200 OK Server: nginx/1.15.2 Tip In order to preserve the source IP address in HTTP requests sent to NGINX, it is necessary to use the Local traffic policy. Traffic policies are described in more details in Traffic policies as well as in the next section.","title":"A pure software solution: MetalLB"},{"location":"deploy/baremetal/#over-a-nodeport-service","text":"Due to its simplicity, this is the setup a user will deploy by default when following the steps described in the installation guide . Info A Service of type NodePort exposes, via the kube-proxy component, the same unprivileged port (default: 30000-32767) on every Kubernetes node, masters included. For more information, see Services . In this configuration, the NGINX container remains isolated from the host network. As a result, it can safely bind to any port, including the standard HTTP ports 80 and 443. However, due to the container namespace isolation, a client located outside the cluster network (e.g. on the public internet) is not able to access Ingress hosts directly on ports 80 and 443. Instead, the external client must append the NodePort allocated to the ingress-nginx Service to HTTP requests. Example Given the NodePort 30100 allocated to the ingress-nginx Service $ kubectl -n ingress-nginx get svc NAME TYPE CLUSTER-IP PORT(S) default-http-backend ClusterIP 10.0.64.249 80/TCP ingress-nginx NodePort 10.0.220.217 80:30100/TCP,443:30101/TCP and a Kubernetes node with the public IP address 203.0.113.2 (the external IP is added as an example, in most bare-metal environments this value is ) $ kubectl get node NAME STATUS ROLES EXTERNAL-IP host-1 Ready master 203.0.113.1 host-2 Ready node 203.0.113.2 host-3 Ready node 203.0.113.3 a client would reach an Ingress with host: myapp.example.com at http://myapp.example.com:30100 , where the myapp.example.com subdomain resolves to the 203.0.113.2 IP address. Impact on the host system While it may sound tempting to reconfigure the NodePort range using the --service-node-port-range API server flag to include unprivileged ports and be able to expose ports 80 and 443, doing so may result in unexpected issues including (but not limited to) the use of ports otherwise reserved to system daemons and the necessity to grant kube-proxy privileges it may otherwise not require. This practice is therefore discouraged . See the other approaches proposed in this page for alternatives. This approach has a few other limitations one ought to be aware of: Source IP address Services of type NodePort perform source address translation by default. This means the source IP of a HTTP request is always the IP address of the Kubernetes node that received the request from the perspective of NGINX. The recommended way to preserve the source IP in a NodePort setup is to set the value of the externalTrafficPolicy field of the ingress-nginx Service spec to Local ( example ). Warning This setting effectively drops packets sent to Kubernetes nodes which are not running any instance of the NGINX Ingress controller. Consider assigning NGINX Pods to specific nodes in order to control on what nodes the NGINX Ingress controller should be scheduled or not scheduled. Example In a Kubernetes cluster composed of 3 nodes (the external IP is added as an example, in most bare-metal environments this value is ) $ kubectl get node NAME STATUS ROLES EXTERNAL-IP host-1 Ready master 203.0.113.1 host-2 Ready node 203.0.113.2 host-3 Ready node 203.0.113.3 with a ingress-nginx-controller Deployment composed of 2 replicas $ kubectl -n ingress-nginx get pod -o wide NAME READY STATUS IP NODE default-http-backend-7c5bc89cc9-p86md 1/1 Running 172.17.1.1 host-2 ingress-nginx-controller-cf9ff8c96-8vvf8 1/1 Running 172.17.0.3 host-3 ingress-nginx-controller-cf9ff8c96-pxsds 1/1 Running 172.17.1.4 host-2 Requests sent to host-2 and host-3 would be forwarded to NGINX and original client's IP would be preserved, while requests to host-1 would get dropped because there is no NGINX replica running on that node. Ingress status Because NodePort Services do not get a LoadBalancerIP assigned by definition, the NGINX Ingress controller does not update the status of Ingress objects it manages . $ kubectl get ingress NAME HOSTS ADDRESS PORTS test-ingress myapp.example.com 80 Despite the fact there is no load balancer providing a public IP address to the NGINX Ingress controller, it is possible to force the status update of all managed Ingress objects by setting the externalIPs field of the ingress-nginx Service. Warning There is more to setting externalIPs than just enabling the NGINX Ingress controller to update the status of Ingress objects. Please read about this option in the Services page of official Kubernetes documentation as well as the section about External IPs in this document for more information. Example Given the following 3-node Kubernetes cluster (the external IP is added as an example, in most bare-metal environments this value is ) $ kubectl get node NAME STATUS ROLES EXTERNAL-IP host-1 Ready master 203.0.113.1 host-2 Ready node 203.0.113.2 host-3 Ready node 203.0.113.3 one could edit the ingress-nginx Service and add the following field to the object spec spec : externalIPs : - 203.0.113.1 - 203.0.113.2 - 203.0.113.3 which would in turn be reflected on Ingress objects as follows: $ kubectl get ingress -o wide NAME HOSTS ADDRESS PORTS test-ingress myapp.example.com 203.0.113.1,203.0.113.2,203.0.113.3 80 Redirects As NGINX is not aware of the port translation operated by the NodePort Service , backend applications are responsible for generating redirect URLs that take into account the URL used by external clients, including the NodePort. Example Redirects generated by NGINX, for instance HTTP to HTTPS or domain to www.domain , are generated without NodePort: $ curl -D- http://myapp.example.com:30100 ` HTTP/1.1 308 Permanent Redirect Server: nginx/1.15.2 Location: https://myapp.example.com/ #-> missing NodePort in HTTPS redirect","title":"Over a NodePort Service"},{"location":"deploy/baremetal/#via-the-host-network","text":"In a setup where there is no external load balancer available but using NodePorts is not an option, one can configure ingress-nginx Pods to use the network of the host they run on instead of a dedicated network namespace. The benefit of this approach is that the NGINX Ingress controller can bind ports 80 and 443 directly to Kubernetes nodes' network interfaces, without the extra network translation imposed by NodePort Services. Note This approach does not leverage any Service object to expose the NGINX Ingress controller. If the ingress-nginx Service exists in the target cluster, it is recommended to delete it . This can be achieved by enabling the hostNetwork option in the Pods' spec. template : spec : hostNetwork : true Security considerations Enabling this option exposes every system daemon to the NGINX Ingress controller on any network interface, including the host's loopback. Please evaluate the impact this may have on the security of your system carefully. Example Consider this ingress-nginx-controller Deployment composed of 2 replicas, NGINX Pods inherit from the IP address of their host instead of an internal Pod IP. $ kubectl -n ingress-nginx get pod -o wide NAME READY STATUS IP NODE default-http-backend-7c5bc89cc9-p86md 1/1 Running 172.17.1.1 host-2 ingress-nginx-controller-5b4cf5fc6-7lg6c 1/1 Running 203.0.113.3 host-3 ingress-nginx-controller-5b4cf5fc6-lzrls 1/1 Running 203.0.113.2 host-2 One major limitation of this deployment approach is that only a single NGINX Ingress controller Pod may be scheduled on each cluster node, because binding the same port multiple times on the same network interface is technically impossible. Pods that are unschedulable due to such situation fail with the following event: $ kubectl -n ingress-nginx describe pod ... Events: Type Reason From Message ---- ------ ---- ------- Warning FailedScheduling default-scheduler 0/3 nodes are available: 3 node(s) didn't have free ports for the requested pod ports. One way to ensure only schedulable Pods are created is to deploy the NGINX Ingress controller as a DaemonSet instead of a traditional Deployment. Info A DaemonSet schedules exactly one type of Pod per cluster node, masters included, unless a node is configured to repel those Pods . For more information, see DaemonSet . Because most properties of DaemonSet objects are identical to Deployment objects, this documentation page leaves the configuration of the corresponding manifest at the user's discretion. Like with NodePorts, this approach has a few quirks it is important to be aware of. DNS resolution Pods configured with hostNetwork: true do not use the internal DNS resolver (i.e. kube-dns or CoreDNS ), unless their dnsPolicy spec field is set to ClusterFirstWithHostNet . Consider using this setting if NGINX is expected to resolve internal names for any reason. Ingress status Because there is no Service exposing the NGINX Ingress controller in a configuration using the host network, the default --publish-service flag used in standard cloud setups does not apply and the status of all Ingress objects remains blank. $ kubectl get ingress NAME HOSTS ADDRESS PORTS test-ingress myapp.example.com 80 Instead, and because bare-metal nodes usually don't have an ExternalIP, one has to enable the --report-node-internal-ip-address flag, which sets the status of all Ingress objects to the internal IP address of all nodes running the NGINX Ingress controller. Example Given a ingress-nginx-controller DaemonSet composed of 2 replicas $ kubectl -n ingress-nginx get pod -o wide NAME READY STATUS IP NODE default-http-backend-7c5bc89cc9-p86md 1/1 Running 172.17.1.1 host-2 ingress-nginx-controller-5b4cf5fc6-7lg6c 1/1 Running 203.0.113.3 host-3 ingress-nginx-controller-5b4cf5fc6-lzrls 1/1 Running 203.0.113.2 host-2 the controller sets the status of all Ingress objects it manages to the following value: $ kubectl get ingress -o wide NAME HOSTS ADDRESS PORTS test-ingress myapp.example.com 203.0.113.2,203.0.113.3 80 Note Alternatively, it is possible to override the address written to Ingress objects using the --publish-status-address flag. See Command line arguments .","title":"Via the host network"},{"location":"deploy/baremetal/#using-a-self-provisioned-edge","text":"Similarly to cloud environments, this deployment approach requires an edge network component providing a public entrypoint to the Kubernetes cluster. This edge component can be either hardware (e.g. vendor appliance) or software (e.g. HAproxy ) and is usually managed outside of the Kubernetes landscape by operations teams. Such deployment builds upon the NodePort Service described above in Over a NodePort Service , with one significant difference: external clients do not access cluster nodes directly, only the edge component does. This is particularly suitable for private Kubernetes clusters where none of the nodes has a public IP address. On the edge side, the only prerequisite is to dedicate a public IP address that forwards all HTTP traffic to Kubernetes nodes and/or masters. Incoming traffic on TCP ports 80 and 443 is forwarded to the corresponding HTTP and HTTPS NodePort on the target nodes as shown in the diagram below:","title":"Using a self-provisioned edge"},{"location":"deploy/baremetal/#external-ips","text":"Source IP address This method does not allow preserving the source IP of HTTP requests in any manner, it is therefore not recommended to use it despite its apparent simplicity. The externalIPs Service option was previously mentioned in the NodePort section. As per the Services page of the official Kubernetes documentation, the externalIPs option causes kube-proxy to route traffic sent to arbitrary IP addresses and on the Service ports to the endpoints of that Service. These IP addresses must belong to the target node . Example Given the following 3-node Kubernetes cluster (the external IP is added as an example, in most bare-metal environments this value is ) $ kubectl get node NAME STATUS ROLES EXTERNAL-IP host-1 Ready master 203.0.113.1 host-2 Ready node 203.0.113.2 host-3 Ready node 203.0.113.3 and the following ingress-nginx NodePort Service $ kubectl -n ingress-nginx get svc NAME TYPE CLUSTER-IP PORT(S) ingress-nginx NodePort 10.0.220.217 80:30100/TCP,443:30101/TCP One could set the following external IPs in the Service spec, and NGINX would become available on both the NodePort and the Service port: spec : externalIPs : - 203.0.113.2 - 203.0.113.3 $ curl -D- http://myapp.example.com:30100 HTTP/1.1 200 OK Server: nginx/1.15.2 $ curl -D- http://myapp.example.com HTTP/1.1 200 OK Server: nginx/1.15.2 We assume the myapp.example.com subdomain above resolves to both 203.0.113.2 and 203.0.113.3 IP addresses.","title":"External IPs"},{"location":"deploy/hardening-guide/","text":"Hardening Guide \u00b6 Overview \u00b6 There are several ways to do hardening and securing of nginx. In this documentation two guides are used, the guides are overlapping in some points: nginx CIS Benchmark cipherlist.eu (one of many forks of the now dead project cipherli.st) This guide describes, what of the different configurations described in those guides is already implemented as default in the nginx implementation of kubernetes ingress, what needs to be configured, what is obsolete due to the fact that the nginx is running as container (the CIS benchmark relates to a non-containerized installation) and what is difficult or not possible. Be aware that this is only a guide and you are responsible for your own implementation. Some of the configurations may lead to have specific clients unable to reach your site or similar consequences. This guide refers to chapters in the CIS Benchmark. For full explanation you should refer to the benchmark document itself Configuration Guide \u00b6 Chapter in CIS benchmark Status Default Action to do if not default 1 Initial Setup 1.1 Installation 1.1.1 Ensure NGINX is installed (Scored) OK done through helm charts / following documentation to deploy nginx ingress 1.1.2 Ensure NGINX is installed from source (Not Scored) OK done through helm charts / following documentation to deploy nginx ingress 1.2 Configure Software Updates 1.2.1 Ensure package manager repositories are properly configured (Not Scored) OK done via helm, nginx version could be overwritten, however compability is not ensured then 1.2.2 Ensure the latest software package is installed (Not Scored) ACTION NEEDED done via helm, nginx version could be overwritten, however compability is not ensured then Plan for periodic updates 2 Basic Configuration 2.1 Minimize NGINX Modules 2.1.1 Ensure only required modules are installed (Not Scored) OK Already only needed modules are installed, however proposals for further reduction are welcome 2.1.2 Ensure HTTP WebDAV module is not installed (Scored) RISK TO BE ACCEPTED It is installed, see compile options here . Disabling that would require building own image for nginx ingress controller. The effort is too high in comparison to the achieved effect 2.1.3 Ensure modules with gzip functionality are disabled (Scored) RISK TO BE ACCEPTED See previous answer 2.1.4 Ensure the autoindex module is disabled (Scored) OK No autoindex configs so far in ingress defaults 2.2 Account Security 2.2.1 Ensure that NGINX is run using a non-privileged, dedicated service account (Not Scored) OK Pod configured as user www-data: See this line in helm chart values . Compiled with user www-data: See this line in build script 2.2.2 Ensure the NGINX service account is locked (Scored) OK Docker design ensures this 2.2.3 Ensure the NGINX service account has an invalid shell (Scored) OK Shell is nologin: see this line in build script 2.3 Permissions and Ownership 2.3.1 Ensure NGINX directories and files are owned by root (Scored) OK Obsolete through docker-design and ingress controller needs to update the configs dynamically 2.3.2 Ensure access to NGINX directories and files is restricted (Scored) OK See previous answer 2.3.3 Ensure the NGINX process ID (PID) file is secured (Scored) OK No PID-File due to docker design 2.3.4 Ensure the core dump directory is secured (Not Scored) OK No working_directory configured by default 2.4 Network Configuration 2.4.1 Ensure NGINX only listens for network connections on authorized ports (Not Scored) OK Ensured by automatic nginx.conf configuration 2.4.2 Ensure requests for unknown host names are rejected (Not Scored) OK They are not rejected but send to the \"default backend\" delivering approriate errors (mostly 404) 2.4.3 Ensure keepalive_timeout is 10 seconds or less, but not 0 (Scored) ACTION NEEDED Default is 75s configure keep-alive to 10 seconds according to this documentation 2.4.4 Ensure send_timeout is set to 10 seconds or less, but not 0 (Scored) RISK TO BE ACCEPTED Not configured, however the nginx default is 60s Not configurable 2.5 Information Disclosure 2.5.1 Ensure server_tokens directive is set to off (Scored) OK server_tokens is configured to off by defaukt 2.5.2 Ensure default error and index.html pages do not reference NGINX (Scored) ACTION NEEDED 404 shows no version at all, 503 and 403 show \"nginx\", which is hardcoded see this line in nginx source code configure custom error pages at least for 403, 404 and 503 and 500 2.5.3 Ensure hidden file serving is disabled (Not Scored) ACTION NEEDED config not set configure a config.server-snippet Snippet, but beware of .well-known challenges or similar. Refer to the benchmark here please 2.5.4 Ensure the NGINX reverse proxy does not enable information disclosure (Scored) ACTION NEEDED hide not configured configure hide-headers with array of \"X-Powered-By\" and \"Server\": according to this documentation 3 Logging 3.1 Ensure detailed logging is enabled (Not Scored) OK Ningx ingress has a very detailled log format by default 3.2 Ensure access logging is enabled (Scored) OK Access log is enabled by default 3.3 Ensure error logging is enabled and set to the info logging level (Scored) OK Error log is configured by default. The log level does not matter, because it is all sent to STDOUT anyway 3.4 Ensure log files are rotated (Scored) OBSOLETE Log file handling is not part of the nginx ingress and should be handled separatly 3.5 Ensure error logs are sent to a remote syslog server (Not Scored) OBSOLETE See previous answer 3.6 Ensure access logs are sent to a remote syslog server (Not Scored) OBSOLETE See previous answer 3.7 Ensure proxies pass source IP information (Scored) OK Headers are set by default 4 Encryption 4.1 TLS / SSL Configuration 4.1.1 Ensure HTTP is redirected to HTTPS (Scored) OK Redirect to TLS is default 4.1.2 Ensure a trusted certificate and trust chain is installed (Not Scored) ACTION NEEDED For installing certs there are enough manuals in the web. A good way is to use lets encrypt through cert-manager Install proper certificates or use lets encrypt with cert-manager 4.1.3 Ensure private key permissions are restricted (Scored) ACTION NEEDED See previous answer 4.1.4 Ensure only modern TLS protocols are used (Scored) OK/ACTION NEEDED Default is TLS 1.2 + 1.3, while this is okay for CIS Benchmark, cipherlist.eu only recommends 1.3. This may cut off old OS's Set controller.config.ssl-protocols to \"TLSv1.3\" 4.1.5 Disable weak ciphers (Scored) ACTION NEEDED Default ciphers are already good, but cipherlist.eu recommends even stronger ciphers Set controller.config.ssl-ciphers to \"EECDH+AESGCM:EDH+AESGCM\" 4.1.6 Ensure custom Diffie-Hellman parameters are used (Scored) ACTION NEEDED No custom DH parameters are generated Generate dh parameters for each ingress deployment you use - see here for a how to 4.1.7 Ensure Online Certificate Status Protocol (OCSP) stapling is enabled (Scored) ACTION NEEDED Not enabled set via this configuration parameter 4.1.8 Ensure HTTP Strict Transport Security (HSTS) is enabled (Scored) OK HSTS is enabled by default 4.1.9 Ensure HTTP Public Key Pinning is enabled (Not Scored) ACTION NEEDED / RISK TO BE ACCEPTED HKPK not enabled by default If lets encrypt is not used, set correct HPKP header. There are several ways to implement this - with the helm charts it works via controller.add-headers. If lets encrypt is used, this is complicated, a solution here is yet unknown 4.1.10 Ensure upstream server traffic is authenticated with a client certificate (Scored) DEPENDS ON BACKEND Highly dependend on backends, not every backend allows configuring this, can also be mitigated via a service mesh If backend allows it, manual is here 4.1.11 Ensure the upstream traffic server certificate is trusted (Not Scored) DEPENDS ON BACKEND Highly dependend on backends, not every backend allows configuring this, can also be mitigated via a service mesh If backend allows it, see configuration here 4.1.12 Ensure your domain is preloaded (Not Scored) ACTION NEEDED Preload is not active by default Set controller.config.hsts-preload to true 4.1.13 Ensure session resumption is disabled to enable perfect forward security (Scored) ACTION NEEDED Session tickets are enabled by default Set controller.config.ssl-session-tickets to false 4.1.14 Ensure HTTP/2.0 is used (Not Scored) OK http2 is set by default 5 Request Filtering and Restrictions 5.1 Access Control 5.1.1 Ensure allow and deny filters limit access to specific IP addresses (Not Scored) OK/ACTION NEEDED Depends on use case, geo ip module is compiled into nginx ingress controller, there are several ways to use it If needed set IP restrictions via annotations or work with config snippets (be careful with lets-encrypt-http-challenge!) 5.1.2 Ensure only whitelisted HTTP methods are allowed (Not Scored) OK/ACTION NEEDED Depends on use case If required it can be set via config snippet 5.2 Request Limits 5.2.1 Ensure timeout values for reading the client header and body are set correctly (Scored) ACTION NEEDED Default timeout is 60s Set via this configuration parameter and respective body aequivalent 5.2.2 Ensure the maximum request body size is set correctly (Scored) ACTION NEEDED Default is 1m set via this configuration parameter 5.2.3 Ensure the maximum buffer size for URIs is defined (Scored) ACTION NEEDED Default is 4 8k Set via this configuration parameter 5.2.4 Ensure the number of connections per IP address is limited (Not Scored) OK/ACTION NEEDED No limit set Depends on use case, limit can be set via these annotations 5.2.5 Ensure rate limits by IP address are set (Not Scored) OK/ACTION NEEDED No limit set Depends on use case, limit can be set via these annotations 5.3 Browser Security 5.3.1 Ensure X-Frame-Options header is configured and enabled (Scored) ACTION NEEDED Header not set by default Several ways to implement this - with the helm charts it works via controller.add-headers 5.3.2 Ensure X-Content-Type-Options header is configured and enabled (Scored) ACTION NEEDED See previous answer See previous answer 5.3.3 Ensure the X-XSS-Protection Header is enabled and configured properly (Scored) ACTION NEEDED See previous answer See previous answer 5.3.4 Ensure that Content Security Policy (CSP) is enabled and configured properly (Not Scored) ACTION NEEDED See previous answer See previous answer 5.3.5 Ensure the Referrer Policy is enabled and configured properly (Not Scored) ACTION NEEDED Depends on application. It should be handled in the applications webserver itself, not in the load balancing ingress check backend webserver 6 Mandatory Access Control n/a too high level, depends on backends @media only screen and (min-width: 768px) { td:nth-child(1){ white-space:normal !important; } .md-typeset table:not([class]) td { padding: .2rem .3rem; } }","title":"Hardening guide"},{"location":"deploy/hardening-guide/#hardening-guide","text":"","title":"Hardening Guide"},{"location":"deploy/hardening-guide/#overview","text":"There are several ways to do hardening and securing of nginx. In this documentation two guides are used, the guides are overlapping in some points: nginx CIS Benchmark cipherlist.eu (one of many forks of the now dead project cipherli.st) This guide describes, what of the different configurations described in those guides is already implemented as default in the nginx implementation of kubernetes ingress, what needs to be configured, what is obsolete due to the fact that the nginx is running as container (the CIS benchmark relates to a non-containerized installation) and what is difficult or not possible. Be aware that this is only a guide and you are responsible for your own implementation. Some of the configurations may lead to have specific clients unable to reach your site or similar consequences. This guide refers to chapters in the CIS Benchmark. For full explanation you should refer to the benchmark document itself","title":"Overview"},{"location":"deploy/hardening-guide/#configuration-guide","text":"Chapter in CIS benchmark Status Default Action to do if not default 1 Initial Setup 1.1 Installation 1.1.1 Ensure NGINX is installed (Scored) OK done through helm charts / following documentation to deploy nginx ingress 1.1.2 Ensure NGINX is installed from source (Not Scored) OK done through helm charts / following documentation to deploy nginx ingress 1.2 Configure Software Updates 1.2.1 Ensure package manager repositories are properly configured (Not Scored) OK done via helm, nginx version could be overwritten, however compability is not ensured then 1.2.2 Ensure the latest software package is installed (Not Scored) ACTION NEEDED done via helm, nginx version could be overwritten, however compability is not ensured then Plan for periodic updates 2 Basic Configuration 2.1 Minimize NGINX Modules 2.1.1 Ensure only required modules are installed (Not Scored) OK Already only needed modules are installed, however proposals for further reduction are welcome 2.1.2 Ensure HTTP WebDAV module is not installed (Scored) RISK TO BE ACCEPTED It is installed, see compile options here . Disabling that would require building own image for nginx ingress controller. The effort is too high in comparison to the achieved effect 2.1.3 Ensure modules with gzip functionality are disabled (Scored) RISK TO BE ACCEPTED See previous answer 2.1.4 Ensure the autoindex module is disabled (Scored) OK No autoindex configs so far in ingress defaults 2.2 Account Security 2.2.1 Ensure that NGINX is run using a non-privileged, dedicated service account (Not Scored) OK Pod configured as user www-data: See this line in helm chart values . Compiled with user www-data: See this line in build script 2.2.2 Ensure the NGINX service account is locked (Scored) OK Docker design ensures this 2.2.3 Ensure the NGINX service account has an invalid shell (Scored) OK Shell is nologin: see this line in build script 2.3 Permissions and Ownership 2.3.1 Ensure NGINX directories and files are owned by root (Scored) OK Obsolete through docker-design and ingress controller needs to update the configs dynamically 2.3.2 Ensure access to NGINX directories and files is restricted (Scored) OK See previous answer 2.3.3 Ensure the NGINX process ID (PID) file is secured (Scored) OK No PID-File due to docker design 2.3.4 Ensure the core dump directory is secured (Not Scored) OK No working_directory configured by default 2.4 Network Configuration 2.4.1 Ensure NGINX only listens for network connections on authorized ports (Not Scored) OK Ensured by automatic nginx.conf configuration 2.4.2 Ensure requests for unknown host names are rejected (Not Scored) OK They are not rejected but send to the \"default backend\" delivering approriate errors (mostly 404) 2.4.3 Ensure keepalive_timeout is 10 seconds or less, but not 0 (Scored) ACTION NEEDED Default is 75s configure keep-alive to 10 seconds according to this documentation 2.4.4 Ensure send_timeout is set to 10 seconds or less, but not 0 (Scored) RISK TO BE ACCEPTED Not configured, however the nginx default is 60s Not configurable 2.5 Information Disclosure 2.5.1 Ensure server_tokens directive is set to off (Scored) OK server_tokens is configured to off by defaukt 2.5.2 Ensure default error and index.html pages do not reference NGINX (Scored) ACTION NEEDED 404 shows no version at all, 503 and 403 show \"nginx\", which is hardcoded see this line in nginx source code configure custom error pages at least for 403, 404 and 503 and 500 2.5.3 Ensure hidden file serving is disabled (Not Scored) ACTION NEEDED config not set configure a config.server-snippet Snippet, but beware of .well-known challenges or similar. Refer to the benchmark here please 2.5.4 Ensure the NGINX reverse proxy does not enable information disclosure (Scored) ACTION NEEDED hide not configured configure hide-headers with array of \"X-Powered-By\" and \"Server\": according to this documentation 3 Logging 3.1 Ensure detailed logging is enabled (Not Scored) OK Ningx ingress has a very detailled log format by default 3.2 Ensure access logging is enabled (Scored) OK Access log is enabled by default 3.3 Ensure error logging is enabled and set to the info logging level (Scored) OK Error log is configured by default. The log level does not matter, because it is all sent to STDOUT anyway 3.4 Ensure log files are rotated (Scored) OBSOLETE Log file handling is not part of the nginx ingress and should be handled separatly 3.5 Ensure error logs are sent to a remote syslog server (Not Scored) OBSOLETE See previous answer 3.6 Ensure access logs are sent to a remote syslog server (Not Scored) OBSOLETE See previous answer 3.7 Ensure proxies pass source IP information (Scored) OK Headers are set by default 4 Encryption 4.1 TLS / SSL Configuration 4.1.1 Ensure HTTP is redirected to HTTPS (Scored) OK Redirect to TLS is default 4.1.2 Ensure a trusted certificate and trust chain is installed (Not Scored) ACTION NEEDED For installing certs there are enough manuals in the web. A good way is to use lets encrypt through cert-manager Install proper certificates or use lets encrypt with cert-manager 4.1.3 Ensure private key permissions are restricted (Scored) ACTION NEEDED See previous answer 4.1.4 Ensure only modern TLS protocols are used (Scored) OK/ACTION NEEDED Default is TLS 1.2 + 1.3, while this is okay for CIS Benchmark, cipherlist.eu only recommends 1.3. This may cut off old OS's Set controller.config.ssl-protocols to \"TLSv1.3\" 4.1.5 Disable weak ciphers (Scored) ACTION NEEDED Default ciphers are already good, but cipherlist.eu recommends even stronger ciphers Set controller.config.ssl-ciphers to \"EECDH+AESGCM:EDH+AESGCM\" 4.1.6 Ensure custom Diffie-Hellman parameters are used (Scored) ACTION NEEDED No custom DH parameters are generated Generate dh parameters for each ingress deployment you use - see here for a how to 4.1.7 Ensure Online Certificate Status Protocol (OCSP) stapling is enabled (Scored) ACTION NEEDED Not enabled set via this configuration parameter 4.1.8 Ensure HTTP Strict Transport Security (HSTS) is enabled (Scored) OK HSTS is enabled by default 4.1.9 Ensure HTTP Public Key Pinning is enabled (Not Scored) ACTION NEEDED / RISK TO BE ACCEPTED HKPK not enabled by default If lets encrypt is not used, set correct HPKP header. There are several ways to implement this - with the helm charts it works via controller.add-headers. If lets encrypt is used, this is complicated, a solution here is yet unknown 4.1.10 Ensure upstream server traffic is authenticated with a client certificate (Scored) DEPENDS ON BACKEND Highly dependend on backends, not every backend allows configuring this, can also be mitigated via a service mesh If backend allows it, manual is here 4.1.11 Ensure the upstream traffic server certificate is trusted (Not Scored) DEPENDS ON BACKEND Highly dependend on backends, not every backend allows configuring this, can also be mitigated via a service mesh If backend allows it, see configuration here 4.1.12 Ensure your domain is preloaded (Not Scored) ACTION NEEDED Preload is not active by default Set controller.config.hsts-preload to true 4.1.13 Ensure session resumption is disabled to enable perfect forward security (Scored) ACTION NEEDED Session tickets are enabled by default Set controller.config.ssl-session-tickets to false 4.1.14 Ensure HTTP/2.0 is used (Not Scored) OK http2 is set by default 5 Request Filtering and Restrictions 5.1 Access Control 5.1.1 Ensure allow and deny filters limit access to specific IP addresses (Not Scored) OK/ACTION NEEDED Depends on use case, geo ip module is compiled into nginx ingress controller, there are several ways to use it If needed set IP restrictions via annotations or work with config snippets (be careful with lets-encrypt-http-challenge!) 5.1.2 Ensure only whitelisted HTTP methods are allowed (Not Scored) OK/ACTION NEEDED Depends on use case If required it can be set via config snippet 5.2 Request Limits 5.2.1 Ensure timeout values for reading the client header and body are set correctly (Scored) ACTION NEEDED Default timeout is 60s Set via this configuration parameter and respective body aequivalent 5.2.2 Ensure the maximum request body size is set correctly (Scored) ACTION NEEDED Default is 1m set via this configuration parameter 5.2.3 Ensure the maximum buffer size for URIs is defined (Scored) ACTION NEEDED Default is 4 8k Set via this configuration parameter 5.2.4 Ensure the number of connections per IP address is limited (Not Scored) OK/ACTION NEEDED No limit set Depends on use case, limit can be set via these annotations 5.2.5 Ensure rate limits by IP address are set (Not Scored) OK/ACTION NEEDED No limit set Depends on use case, limit can be set via these annotations 5.3 Browser Security 5.3.1 Ensure X-Frame-Options header is configured and enabled (Scored) ACTION NEEDED Header not set by default Several ways to implement this - with the helm charts it works via controller.add-headers 5.3.2 Ensure X-Content-Type-Options header is configured and enabled (Scored) ACTION NEEDED See previous answer See previous answer 5.3.3 Ensure the X-XSS-Protection Header is enabled and configured properly (Scored) ACTION NEEDED See previous answer See previous answer 5.3.4 Ensure that Content Security Policy (CSP) is enabled and configured properly (Not Scored) ACTION NEEDED See previous answer See previous answer 5.3.5 Ensure the Referrer Policy is enabled and configured properly (Not Scored) ACTION NEEDED Depends on application. It should be handled in the applications webserver itself, not in the load balancing ingress check backend webserver 6 Mandatory Access Control n/a too high level, depends on backends @media only screen and (min-width: 768px) { td:nth-child(1){ white-space:normal !important; } .md-typeset table:not([class]) td { padding: .2rem .3rem; } }","title":"Configuration Guide"},{"location":"deploy/rbac/","text":"Role Based Access Control (RBAC) \u00b6 Overview \u00b6 This example applies to nginx-ingress-controllers being deployed in an environment with RBAC enabled. Role Based Access Control is comprised of four layers: ClusterRole - permissions assigned to a role that apply to an entire cluster ClusterRoleBinding - binding a ClusterRole to a specific account Role - permissions assigned to a role that apply to a specific namespace RoleBinding - binding a Role to a specific account In order for RBAC to be applied to an nginx-ingress-controller, that controller should be assigned to a ServiceAccount . That ServiceAccount should be bound to the Role s and ClusterRole s defined for the nginx-ingress-controller. Service Accounts created in this example \u00b6 One ServiceAccount is created in this example, nginx-ingress-serviceaccount . Permissions Granted in this example \u00b6 There are two sets of permissions defined in this example. Cluster-wide permissions defined by the ClusterRole named nginx-ingress-clusterrole , and namespace specific permissions defined by the Role named nginx-ingress-role . Cluster Permissions \u00b6 These permissions are granted in order for the nginx-ingress-controller to be able to function as an ingress across the cluster. These permissions are granted to the ClusterRole named nginx-ingress-clusterrole configmaps , endpoints , nodes , pods , secrets : list, watch nodes : get services , ingresses : get, list, watch events : create, patch ingresses/status : update Namespace Permissions \u00b6 These permissions are granted specific to the nginx-ingress namespace. These permissions are granted to the Role named nginx-ingress-role configmaps , pods , secrets : get endpoints : get Furthermore to support leader-election, the nginx-ingress-controller needs to have access to a configmap using the resourceName ingress-controller-leader-nginx Note that resourceNames can NOT be used to limit requests using the \u201ccreate\u201d verb because authorizers only have access to information that can be obtained from the request URL, method, and headers (resource names in a \u201ccreate\u201d request are part of the request body). configmaps : get, update (for resourceName ingress-controller-leader-nginx ) configmaps : create This resourceName is the concatenation of the election-id and the ingress-class as defined by the ingress-controller, which defaults to: election-id : ingress-controller-leader ingress-class : nginx resourceName : - Please adapt accordingly if you overwrite either parameter when launching the nginx-ingress-controller. Bindings \u00b6 The ServiceAccount nginx-ingress-serviceaccount is bound to the Role nginx-ingress-role and the ClusterRole nginx-ingress-clusterrole . The serviceAccountName associated with the containers in the deployment must match the serviceAccount. The namespace references in the Deployment metadata, container arguments, and POD_NAMESPACE should be in the nginx-ingress namespace.","title":"Role Based Access Control (RBAC)"},{"location":"deploy/rbac/#role-based-access-control-rbac","text":"","title":"Role Based Access Control (RBAC)"},{"location":"deploy/rbac/#overview","text":"This example applies to nginx-ingress-controllers being deployed in an environment with RBAC enabled. Role Based Access Control is comprised of four layers: ClusterRole - permissions assigned to a role that apply to an entire cluster ClusterRoleBinding - binding a ClusterRole to a specific account Role - permissions assigned to a role that apply to a specific namespace RoleBinding - binding a Role to a specific account In order for RBAC to be applied to an nginx-ingress-controller, that controller should be assigned to a ServiceAccount . That ServiceAccount should be bound to the Role s and ClusterRole s defined for the nginx-ingress-controller.","title":"Overview"},{"location":"deploy/rbac/#service-accounts-created-in-this-example","text":"One ServiceAccount is created in this example, nginx-ingress-serviceaccount .","title":"Service Accounts created in this example"},{"location":"deploy/rbac/#permissions-granted-in-this-example","text":"There are two sets of permissions defined in this example. Cluster-wide permissions defined by the ClusterRole named nginx-ingress-clusterrole , and namespace specific permissions defined by the Role named nginx-ingress-role .","title":"Permissions Granted in this example"},{"location":"deploy/rbac/#cluster-permissions","text":"These permissions are granted in order for the nginx-ingress-controller to be able to function as an ingress across the cluster. These permissions are granted to the ClusterRole named nginx-ingress-clusterrole configmaps , endpoints , nodes , pods , secrets : list, watch nodes : get services , ingresses : get, list, watch events : create, patch ingresses/status : update","title":"Cluster Permissions"},{"location":"deploy/rbac/#namespace-permissions","text":"These permissions are granted specific to the nginx-ingress namespace. These permissions are granted to the Role named nginx-ingress-role configmaps , pods , secrets : get endpoints : get Furthermore to support leader-election, the nginx-ingress-controller needs to have access to a configmap using the resourceName ingress-controller-leader-nginx Note that resourceNames can NOT be used to limit requests using the \u201ccreate\u201d verb because authorizers only have access to information that can be obtained from the request URL, method, and headers (resource names in a \u201ccreate\u201d request are part of the request body). configmaps : get, update (for resourceName ingress-controller-leader-nginx ) configmaps : create This resourceName is the concatenation of the election-id and the ingress-class as defined by the ingress-controller, which defaults to: election-id : ingress-controller-leader ingress-class : nginx resourceName : - Please adapt accordingly if you overwrite either parameter when launching the nginx-ingress-controller.","title":"Namespace Permissions"},{"location":"deploy/rbac/#bindings","text":"The ServiceAccount nginx-ingress-serviceaccount is bound to the Role nginx-ingress-role and the ClusterRole nginx-ingress-clusterrole . The serviceAccountName associated with the containers in the deployment must match the serviceAccount. The namespace references in the Deployment metadata, container arguments, and POD_NAMESPACE should be in the nginx-ingress namespace.","title":"Bindings"},{"location":"deploy/upgrade/","text":"Upgrading \u00b6 Important No matter the method you use for upgrading, if you use template overrides, make sure your templates are compatible with the new version of ingress-nginx . Without Helm \u00b6 To upgrade your ingress-nginx installation, it should be enough to change the version of the image in the controller Deployment. I.e. if your deployment resource looks like (partial example): kind : Deployment metadata : name : nginx-ingress-controller namespace : ingress-nginx spec : replicas : 1 selector : ... template : metadata : ... spec : containers : - name : nginx-ingress-controller image : k8s.gcr.io/ingress-nginx/controller:v0.34.0@sha256:56633bd00dab33d92ba14c6e709126a762d54a75a6e72437adefeaaca0abb069 args : ... simply change the 0.34.0 tag to the version you wish to upgrade to. The easiest way to do this is e.g. (do note you may need to change the name parameter according to your installation): kubectl set image deployment/nginx-ingress-controller \\ nginx-ingress-controller=k8s.gcr.io/ingress-nginx/controller:v0.34.1@sha256:0e072dddd1f7f8fc8909a2ca6f65e76c5f0d2fcfb8be47935ae3457e8bbceb20 \\ -n ingress-nginx For interactive editing, use kubectl edit deployment nginx-ingress-controller -n ingress-nginx . With Helm \u00b6 If you installed ingress-nginx using the Helm command in the deployment docs so its name is ngx-ingress , you should be able to upgrade using helm upgrade --reuse-values ngx-ingress ingress-nginx/ingress-nginx Migrating from stable/nginx-ingress \u00b6 See detailed steps in the upgrading section of the ingress-nginx chart README .","title":"Upgrade"},{"location":"deploy/upgrade/#upgrading","text":"Important No matter the method you use for upgrading, if you use template overrides, make sure your templates are compatible with the new version of ingress-nginx .","title":"Upgrading"},{"location":"deploy/upgrade/#without-helm","text":"To upgrade your ingress-nginx installation, it should be enough to change the version of the image in the controller Deployment. I.e. if your deployment resource looks like (partial example): kind : Deployment metadata : name : nginx-ingress-controller namespace : ingress-nginx spec : replicas : 1 selector : ... template : metadata : ... spec : containers : - name : nginx-ingress-controller image : k8s.gcr.io/ingress-nginx/controller:v0.34.0@sha256:56633bd00dab33d92ba14c6e709126a762d54a75a6e72437adefeaaca0abb069 args : ... simply change the 0.34.0 tag to the version you wish to upgrade to. The easiest way to do this is e.g. (do note you may need to change the name parameter according to your installation): kubectl set image deployment/nginx-ingress-controller \\ nginx-ingress-controller=k8s.gcr.io/ingress-nginx/controller:v0.34.1@sha256:0e072dddd1f7f8fc8909a2ca6f65e76c5f0d2fcfb8be47935ae3457e8bbceb20 \\ -n ingress-nginx For interactive editing, use kubectl edit deployment nginx-ingress-controller -n ingress-nginx .","title":"Without Helm"},{"location":"deploy/upgrade/#with-helm","text":"If you installed ingress-nginx using the Helm command in the deployment docs so its name is ngx-ingress , you should be able to upgrade using helm upgrade --reuse-values ngx-ingress ingress-nginx/ingress-nginx","title":"With Helm"},{"location":"deploy/upgrade/#migrating-from-stablenginx-ingress","text":"See detailed steps in the upgrading section of the ingress-nginx chart README .","title":"Migrating from stable/nginx-ingress"},{"location":"enhancements/","text":"Kubernetes Enhancement Proposals (KEPs) \u00b6 A Kubernetes Enhancement Proposal (KEP) is a way to propose, communicate and coordinate on new efforts for the Kubernetes project. For this reason, the ingress-nginx project is adopting it. Quick start for the KEP process \u00b6 Follow the process outlined in the KEP template Do I have to use the KEP process? \u00b6 No... but we hope that you will. Over time having a rich set of KEPs in one place will make it easier for people to track what is going on in the community and find a structured historic record. KEPs are only required when the changes are wide ranging and impact most of the project. Why would I want to use the KEP process? \u00b6 Our aim with KEPs is to clearly communicate new efforts to the Kubernetes contributor community. As such, we want to build a well curated set of clear proposals in a common format with useful metadata. Benefits to KEP users (in the limit): Exposure on a kubernetes blessed web site that is findable via web search engines. Cross indexing of KEPs so that users can find connections and the current status of any KEP. A clear process with approvers and reviewers for making decisions. This will lead to more structured decisions that stick as there is a discoverable record around the decisions. We are inspired by IETF RFCs, Python PEPs, and Rust RFCs.","title":"Kubernetes Enhancement Proposals (KEPs)"},{"location":"enhancements/#kubernetes-enhancement-proposals-keps","text":"A Kubernetes Enhancement Proposal (KEP) is a way to propose, communicate and coordinate on new efforts for the Kubernetes project. For this reason, the ingress-nginx project is adopting it.","title":"Kubernetes Enhancement Proposals (KEPs)"},{"location":"enhancements/#quick-start-for-the-kep-process","text":"Follow the process outlined in the KEP template","title":"Quick start for the KEP process"},{"location":"enhancements/#do-i-have-to-use-the-kep-process","text":"No... but we hope that you will. Over time having a rich set of KEPs in one place will make it easier for people to track what is going on in the community and find a structured historic record. KEPs are only required when the changes are wide ranging and impact most of the project.","title":"Do I have to use the KEP process?"},{"location":"enhancements/#why-would-i-want-to-use-the-kep-process","text":"Our aim with KEPs is to clearly communicate new efforts to the Kubernetes contributor community. As such, we want to build a well curated set of clear proposals in a common format with useful metadata. Benefits to KEP users (in the limit): Exposure on a kubernetes blessed web site that is findable via web search engines. Cross indexing of KEPs so that users can find connections and the current status of any KEP. A clear process with approvers and reviewers for making decisions. This will lead to more structured decisions that stick as there is a discoverable record around the decisions. We are inspired by IETF RFCs, Python PEPs, and Rust RFCs.","title":"Why would I want to use the KEP process?"},{"location":"enhancements/20190724-only-dynamic-ssl/","text":"Remove static SSL configuration mode \u00b6 Table of Contents \u00b6 Summary Motivation Goals Non-Goals Proposal Implementation Details/Notes/Constraints Drawbacks Alternatives Summary \u00b6 Since release 0.19.0 is possible to configure SSL certificates without the need of NGINX reloads (thanks to lua) and after release 0.24.0 the default enabled mode is dynamic. Motivation \u00b6 The static configuration implies reloads, something that affects the majority of the users. Goals \u00b6 Deprecation of the flag --enable-dynamic-certificates . Cleanup of the codebase. Non-Goals \u00b6 Features related to certificate authentication are not changed in any way. Proposal \u00b6 Remove static SSL configuration Implementation Details/Notes/Constraints \u00b6 Deprecate the flag Move the directives ssl_certificate and ssl_certificate_key from each server block to the http section. These settings are required to avoid NGINX errors in the logs. Remove any action of the flag --enable-dynamic-certificates Drawbacks \u00b6 Alternatives \u00b6 Keep both implementations","title":"Remove static SSL configuration mode"},{"location":"enhancements/20190724-only-dynamic-ssl/#remove-static-ssl-configuration-mode","text":"","title":"Remove static SSL configuration mode"},{"location":"enhancements/20190724-only-dynamic-ssl/#table-of-contents","text":"Summary Motivation Goals Non-Goals Proposal Implementation Details/Notes/Constraints Drawbacks Alternatives","title":"Table of Contents"},{"location":"enhancements/20190724-only-dynamic-ssl/#summary","text":"Since release 0.19.0 is possible to configure SSL certificates without the need of NGINX reloads (thanks to lua) and after release 0.24.0 the default enabled mode is dynamic.","title":"Summary"},{"location":"enhancements/20190724-only-dynamic-ssl/#motivation","text":"The static configuration implies reloads, something that affects the majority of the users.","title":"Motivation"},{"location":"enhancements/20190724-only-dynamic-ssl/#goals","text":"Deprecation of the flag --enable-dynamic-certificates . Cleanup of the codebase.","title":"Goals"},{"location":"enhancements/20190724-only-dynamic-ssl/#non-goals","text":"Features related to certificate authentication are not changed in any way.","title":"Non-Goals"},{"location":"enhancements/20190724-only-dynamic-ssl/#proposal","text":"Remove static SSL configuration","title":"Proposal"},{"location":"enhancements/20190724-only-dynamic-ssl/#implementation-detailsnotesconstraints","text":"Deprecate the flag Move the directives ssl_certificate and ssl_certificate_key from each server block to the http section. These settings are required to avoid NGINX errors in the logs. Remove any action of the flag --enable-dynamic-certificates","title":"Implementation Details/Notes/Constraints"},{"location":"enhancements/20190724-only-dynamic-ssl/#drawbacks","text":"","title":"Drawbacks"},{"location":"enhancements/20190724-only-dynamic-ssl/#alternatives","text":"Keep both implementations","title":"Alternatives"},{"location":"enhancements/20190815-zone-aware-routing/","text":"Availability zone aware routing \u00b6 Table of Contents \u00b6 Summary Motivation Goals Non-Goals Proposal Implementation History Drawbacks [optional] Summary \u00b6 Teach ingress-nginx about availability zones where endpoints are running in. This way ingress-nginx pod will do its best to proxy to zone-local endpoint. Motivation \u00b6 When users run their services across multiple availability zones they usually pay for egress traffic between zones. Providers such as GCP, Amazon EC charges money for that. ingress-nginx when picking an endpoint to route request to does not consider whether the endpoint is in different zone or the same one. That means it's at least equally likely that it will pick an endpoint from another zone and proxy the request to it. In this situation response from the endpoint to ingress-nginx pod is considered as inter zone traffic and costs money. At the time of this writing GCP charges $0.01 per GB of inter zone egress traffic according to https://cloud.google.com/compute/network-pricing. According to https://datapath.io/resources/blog/what-are-aws-data-transfer-costs-and-how-to-minimize-them/ Amazon also charges the same amount of money sa GCP for cross zone, egress traffic. This can be a lot of money depending on once's traffic. By teaching ingress-nginx about zones we can eliminate or at least decrease this cost. Arguably inter-zone network latency should also be better than cross zone. Goals \u00b6 Given a regional cluster running ingress-nginx, ingress-nginx should do best effort to pick zone-local endpoint when proxying This should not impact canary feature ingress-nginx should be able to operate successfully if there's no zonal endpoints Non-Goals \u00b6 This feature inherently assumes that endpoints are distributed across zones in a way that they can handle all the traffic from ingress-nginx pod(s) in that zone This feature will be relying on https://kubernetes.io/docs/reference/kubernetes-api/labels-annotations-taints/#failure-domainbetakubernetesiozone, it is not this KEP's goal to support other cases Proposal \u00b6 The idea here is to have controller part of ingress-nginx to (1) detect what zone its current pod is running in and (2) detect the zone for every endpoints it knows about. After that it will post that data as part of endpoints to Lua land. Then Lua balancer when picking an endpoint will try to pick zone-local endpoint first and if there is no zone-local endpoint then it will fallback to current behaviour. This feature at least in the beginning should be optional since it is going to make it harder to reason about the load balancing and not everyone might want that. How does controller know what zone it runs in? We can have the pod spec do pass node name using downward API as an environment variable. Then on start controller can get node details from the API based on node name. Once the node details is obtained we can extract the zone from failure-domain.beta.kubernetes.io/zone annotation. Then we can pass that value to Lua land through Nginx configuration when loading lua_ingress.lua module in init_by_lua phase. How do we extract zones for endpoints? We can have the controller watch create and update events on nodes in the entire cluster and based on that keep the map of nodes to zones in the memory. And when we generate endpoints list, we can access node name using .subsets.addresses[i].nodeName and based on that fetch zone from the map in memory and store it as a field on the endpoint. This solution assumes failure-domain.beta.kubernetes.io/zone annotation does not change until the end of node's life. Otherwise we have to watch update events as well on the nodes and that'll add even more overhead. Alternatively, we can get the list of nodes only when there's no node in the memory for given node name. This is probably a better solution because then we would avoid watching for API changes on node resources. We can eagrly fetch all the nodes and build node name to zone mapping on start. And from thereon sync it during endpoints building in the main event loop iff there's no entry exist for the node of an endpoint. This means an extra API call in case cluster has expanded. How do we make sure we do our best to choose zone-local endpoint? This will be done on Lua side. For every backend we will initialize two balancer instances: (1) with all endpoints (2) with all endpoints corresponding to current zone for the backend. Then given the request once we choose what backend needs to serve the request, we will first try to use zonal balancer for that backend. If zonal balancer does not exist (i.e there's no zonal endpoint) then we will use general balancer. In case of zonal outages we assume that readiness probe will fail and controller will see no endpoints for the backend and therefore we will use general balancer. We can enable the feature using a configmap setting. Doing it this way makes it easier to rollback in case of a problem. Implementation History \u00b6 initial version of KEP is shipped proposal and implementation details is done Drawbacks [optional] \u00b6 More load on the Kubernetes API server.","title":"Availability zone aware routing"},{"location":"enhancements/20190815-zone-aware-routing/#availability-zone-aware-routing","text":"","title":"Availability zone aware routing"},{"location":"enhancements/20190815-zone-aware-routing/#table-of-contents","text":"Summary Motivation Goals Non-Goals Proposal Implementation History Drawbacks [optional]","title":"Table of Contents"},{"location":"enhancements/20190815-zone-aware-routing/#summary","text":"Teach ingress-nginx about availability zones where endpoints are running in. This way ingress-nginx pod will do its best to proxy to zone-local endpoint.","title":"Summary"},{"location":"enhancements/20190815-zone-aware-routing/#motivation","text":"When users run their services across multiple availability zones they usually pay for egress traffic between zones. Providers such as GCP, Amazon EC charges money for that. ingress-nginx when picking an endpoint to route request to does not consider whether the endpoint is in different zone or the same one. That means it's at least equally likely that it will pick an endpoint from another zone and proxy the request to it. In this situation response from the endpoint to ingress-nginx pod is considered as inter zone traffic and costs money. At the time of this writing GCP charges $0.01 per GB of inter zone egress traffic according to https://cloud.google.com/compute/network-pricing. According to https://datapath.io/resources/blog/what-are-aws-data-transfer-costs-and-how-to-minimize-them/ Amazon also charges the same amount of money sa GCP for cross zone, egress traffic. This can be a lot of money depending on once's traffic. By teaching ingress-nginx about zones we can eliminate or at least decrease this cost. Arguably inter-zone network latency should also be better than cross zone.","title":"Motivation"},{"location":"enhancements/20190815-zone-aware-routing/#goals","text":"Given a regional cluster running ingress-nginx, ingress-nginx should do best effort to pick zone-local endpoint when proxying This should not impact canary feature ingress-nginx should be able to operate successfully if there's no zonal endpoints","title":"Goals"},{"location":"enhancements/20190815-zone-aware-routing/#non-goals","text":"This feature inherently assumes that endpoints are distributed across zones in a way that they can handle all the traffic from ingress-nginx pod(s) in that zone This feature will be relying on https://kubernetes.io/docs/reference/kubernetes-api/labels-annotations-taints/#failure-domainbetakubernetesiozone, it is not this KEP's goal to support other cases","title":"Non-Goals"},{"location":"enhancements/20190815-zone-aware-routing/#proposal","text":"The idea here is to have controller part of ingress-nginx to (1) detect what zone its current pod is running in and (2) detect the zone for every endpoints it knows about. After that it will post that data as part of endpoints to Lua land. Then Lua balancer when picking an endpoint will try to pick zone-local endpoint first and if there is no zone-local endpoint then it will fallback to current behaviour. This feature at least in the beginning should be optional since it is going to make it harder to reason about the load balancing and not everyone might want that. How does controller know what zone it runs in? We can have the pod spec do pass node name using downward API as an environment variable. Then on start controller can get node details from the API based on node name. Once the node details is obtained we can extract the zone from failure-domain.beta.kubernetes.io/zone annotation. Then we can pass that value to Lua land through Nginx configuration when loading lua_ingress.lua module in init_by_lua phase. How do we extract zones for endpoints? We can have the controller watch create and update events on nodes in the entire cluster and based on that keep the map of nodes to zones in the memory. And when we generate endpoints list, we can access node name using .subsets.addresses[i].nodeName and based on that fetch zone from the map in memory and store it as a field on the endpoint. This solution assumes failure-domain.beta.kubernetes.io/zone annotation does not change until the end of node's life. Otherwise we have to watch update events as well on the nodes and that'll add even more overhead. Alternatively, we can get the list of nodes only when there's no node in the memory for given node name. This is probably a better solution because then we would avoid watching for API changes on node resources. We can eagrly fetch all the nodes and build node name to zone mapping on start. And from thereon sync it during endpoints building in the main event loop iff there's no entry exist for the node of an endpoint. This means an extra API call in case cluster has expanded. How do we make sure we do our best to choose zone-local endpoint? This will be done on Lua side. For every backend we will initialize two balancer instances: (1) with all endpoints (2) with all endpoints corresponding to current zone for the backend. Then given the request once we choose what backend needs to serve the request, we will first try to use zonal balancer for that backend. If zonal balancer does not exist (i.e there's no zonal endpoint) then we will use general balancer. In case of zonal outages we assume that readiness probe will fail and controller will see no endpoints for the backend and therefore we will use general balancer. We can enable the feature using a configmap setting. Doing it this way makes it easier to rollback in case of a problem.","title":"Proposal"},{"location":"enhancements/20190815-zone-aware-routing/#implementation-history","text":"initial version of KEP is shipped proposal and implementation details is done","title":"Implementation History"},{"location":"enhancements/20190815-zone-aware-routing/#drawbacks-optional","text":"More load on the Kubernetes API server.","title":"Drawbacks [optional]"},{"location":"enhancements/YYYYMMDD-kep-template/","text":"Title \u00b6 This is the title of the KEP. Keep it simple and descriptive. A good title can help communicate what the KEP is and should be considered as part of any review. The title should be lowercased and spaces/punctuation should be replaced with - . To get started with this template: Make a copy of this template. Create a copy of this template and name it YYYYMMDD-my-title.md , where YYYYMMDD is the date the KEP was first drafted. Fill out the \"overview\" sections. This includes the Summary and Motivation sections. These should be easy if you've preflighted the idea of the KEP in an issue. Create a PR. Assign it to folks that are sponsoring this process. Create an issue When filing an enhancement tracking issue, please ensure to complete all fields in the template. Merge early. Avoid getting hung up on specific details and instead aim to get the goal of the KEP merged quickly. The best way to do this is to just start with the \"Overview\" sections and fill out details incrementally in follow on PRs. View anything marked as a provisional as a working document and subject to change. Aim for single topic PRs to keep discussions focused. If you disagree with what is already in a document, open a new PR with suggested changes. The canonical place for the latest set of instructions (and the likely source of this file) is here . The Metadata section above is intended to support the creation of tooling around the KEP process. This will be a YAML section that is fenced as a code block. See the KEP process for details on each of these items. Table of Contents \u00b6 A table of contents is helpful for quickly jumping to sections of a KEP and for highlighting any additional information provided beyond the standard KEP template. Ensure the TOC is wrapped with