nginx validate webhook fails some times
When the helm-operator
installs everything on the first run (and during many CI runs), it often fails installing an app with the following error in kubectl describe helmrelease <failed_release>
:
synchronization of release 'prometheus-stack' in namespace 'oas' failed: installation failed: Internal error occurred: failed calling webhook "validate.nginx.ingress.kubernetes.io": Post "https://ingress-ingress-nginx-controller-admission.oas.svc:443/networking/v1beta1/ingresses?timeout=10s": dial tcp 10.43.31.18:443: connect: connection refused
This is most likely due to the fact that flux/helm-operator does not install Nginx before installing the other applications, while those other applications are dependant on the ingress to set up their
When this happens, the helm release fails, gets purged, and usually succeeds on the next helm-operator sync run. Currently there's 20 minutes of waiting time between helm-operator runs, because helm-operator takes up too many resources otherwise...