* Add validation to all annotations
* Add annotation validation for fcgi
* Fix reviews and fcgi e2e
* Add flag to disable cross namespace validation
* Add risk, flag for validation, tests
* Add missing formating
* Enable validation by default on tests
* Test validation flag
* remove ajp from list
* Finalize validation changes
* Add validations to CI
* Update helm docs
* Fix code review
* Use a better name for annotation risk
* datadog: sample_rate omitted by default
* config: use *float32 with nil instead of float32 with sentinel value
* change some names
* gofmt -s -w internal/ingress/controller/nginx.go
* Remove variables with $ before feeding into url.Parse
Signed-off-by: Gerald Pape <gerald@giantswarm.io>
* Do not render invalid request mirroring config
Signed-off-by: Gerald Pape <gerald@giantswarm.io>
* Remove additional note from docs again
Signed-off-by: Gerald Pape <gerald@giantswarm.io>
* Include quotes in e2e test for mirror proxy_pass
---------
Signed-off-by: Gerald Pape <gerald@giantswarm.io>
We can use alternative functions to avoid unnecessary byte/string
conversion calls and reduce allocations.
Signed-off-by: Eng Zer Jun <engzerjun@gmail.com>
* exclude creation and exporting of socket metrics via flag
* make exclude metric naming more consistent
* fix connect time metric update
* add documentation
* e2e test
* improve creation of metric mapping
* Rework Ginkgo usage
Currently Ginkgo is launched multiple times with different options to
accomodate various use-cases. In particular, some specs needs to be run
sequentially because non-namespaced objects are created that conflicts
with concurent Helm deployments.
However Ginkgo is able to handle such cases natively, in particular
specs that needs to be run sequentially are supported (Serial spec).
This commit marks the specs that needs to be run sequentially as Serial
specs and runs the whole test suite from a single Ginkgo invocation. As
a result, a single JUnit report is now generated.
Signed-off-by: Hervé Werner <dud225@hotmail.com>
* Fix controller error in test
Error getting ConfigMap "$NAMESPACE/tcp-services": no object matching key "$NAMESPACE/tcp-services" in local store
Signed-off-by: Hervé Werner <dud225@hotmail.com>
* Replace "go get" invocations by "go install"
Executing "go get" changes the go.mod & go.sum files which is not the
case of "go install".
Signed-off-by: Hervé Werner <dud225@hotmail.com>
* Always clean out the Helm deployment
Signed-off-by: Hervé Werner <dud225@hotmail.com>
* Add E2E test to verify that changes to one or more configmap trigger an update
Signed-off-by: Hervé Werner <dud225@hotmail.com>
---------
Signed-off-by: Hervé Werner <dud225@hotmail.com>
* update path type validation to be false and update e2e test scripts
Signed-off-by: James Strong <strong.james.e@gmail.com>
* update to make tests clear
Signed-off-by: James Strong <strong.james.e@gmail.com>
* update test params
Signed-off-by: James Strong <strong.james.e@gmail.com>
* Adding else per pr comments
Signed-off-by: James Strong <james.strong@chainguard.dev>
---------
Signed-off-by: James Strong <strong.james.e@gmail.com>
Signed-off-by: James Strong <james.strong@chainguard.dev>
* deps: bump k8s dependencies to remove go-autorest
* fix: update use of apiv1.LoadBalancerIngress
Due to changes in the Kubernetes API, we needed to switch to using
v1.IngressLoadBalancerIngress instead of apiv1.LoadBalancerIngress. The
struct is otherwise identical despite the name change.
* fix ingress status test cases
Signed-off-by: Jintao Zhang <zhangjintao9020@gmail.com>
Signed-off-by: Ismayil Mirzali <ismayilmirzeli@gmail.com>
Signed-off-by: Jintao Zhang <zhangjintao9020@gmail.com>
Signed-off-by: Ismayil Mirzali <ismayilmirzeli@gmail.com>
Co-authored-by: Jintao Zhang <zhangjintao9020@gmail.com>
* feat: Add support for IP Deny List
* fixed gomod
* Update package
* go mod tidy
* Revert "go mod tidy"
This reverts commit e6a837e1e7.
* update ginko version
* Updates e2e tests
* fix test typo
