ci: Set hubble.relay.retryTimeout=5s

[chancez]

I noticed CI seemed to flake on hubble-relay being ready. Based on the logs it was due to it being unable to connect to its peers.

Based on the sysdump the endpoints did eventually come up, so I think it just needs to retry more often, and the default is 30s so lower it to 5s in CI.

An example of such a CI run: https://github.com/cilium/cilium/actions/runs/8729875690/job/23952676116?pr=31973.

I've added some temporary commits to test the changes, which I'll remove after I verify the correct options are being used.

[chancez]

Run cilium install --chart-directory=install/kubernetes/cilium --helm-set=debug.enabled=true --helm-set=bpf.monitorAggregation=none --helm-set=dnsProxy.proxyResponseMaxDelay=250ms --helm-set=hubble.relay.retryTimeout=5s --helm-set=image.repository=quay.io/cilium/cilium-ci --helm-set=image.useDigest=false --helm-set=image.tag=a49b5e5d68d4e233ba37117eb7b3e0f0ac1a0a5a --helm-set=operator.image.repository=quay.io/cilium/operator --helm-set=operator.image.suffix=-ci --helm-set=operator.image.tag=a49b5e5d68d4e233ba37117eb7b3e0f0ac1a0a5a --helm-set=operator.image.useDigest=false --helm-set=clustermesh.apiserver.image.repository=quay.io/cilium/clustermesh-apiserver-ci --helm-set=clustermesh.apiserver.image.tag=a49b5e5d68d4e233ba37117eb7b3e0f0ac1a0a5a --helm-set=clustermesh.apiserver.image.useDigest=false --helm-set=hubble.relay.image.repository=quay.io/cilium/hubble-relay-ci --helm-set=hubble.relay.image.tag=a49b5e5d68d4e233ba37117eb7b3e0f0ac1a0a5a --helm-set=hubble.relay.image.useDigest=false --helm-set=hubble.relay.enabled=true --helm-set=cni.chainingMode=portmap --helm-set-string=kubeProxyReplacement=true --helm-set=loadBalancer.l7.backend=envoy --helm-set=tls.secretsBackend=k8s --helm-set=envoy.enabled=false --wait=false

I can see it specified the right option but the test succeeded and there's no sysdumps; so I can't actually confirm it was added to the configmap, but it should be.

[chancez]

Removing the tmp commit so this can be marked as ready for review.

[chancez]

/test

[julianwiedmann]

@chancez should we also backport to stable, to prevent similar CI flakes there?

[chancez]

@julianwiedmann seems reasonable and easy? Do we just apply the necessary labels for that?

[julianwiedmann]

@julianwiedmann seems reasonable and easy? Do we just apply the necessary labels for that?

yep. Looks like the helm-default action doesn't exist in older releases, so leaving it up to you whether we want manual backports for those as well :).