Argo server & controller readiness/liveness check

[decipher27]

Could you let me know why we don't have "Readiness probe" in argo controller? and "liveness" in Argo server?

[tooptoop4]

did u solve? @decipher27

[dixneuf19]

Hi, same question here . There is no official health endpoint for argo workflow server, and in the official Helm chart there is only a readiness probe, but no liveness (source.

First it is a good practice to have both in a Kubernetes app, but more importantly I already had an issue regarding a stuck workflow server, which was fixed with a simple reboot, something a liveness probe would have catch.

Is there any reason to not implement this kind of feature ? Argocd-server does implement this "health" route, and use it in its Helm chart.
If that's not a priority, can we simply use the root / as a custom liveness probe (this is what the chart already does for the readiness) ?