Liveness probe

Hello,

We deploy the camunda-platform helm chart 8.2.8. We noticed that the readiness probe is enabled by default but the liveness probe is not. Just curious, why the liveness probe is not enabled by default? Anything we need to be aware of if we choose to enable it? We would like to enable the liveness probe for both camunda-zeebe-gateway and camunda-zeebe.

We noticed in our k8s cluster that although the gateway and zeebe pods show Running they do not respond and we see the ‘GRPC usage exception – channel closed’ in our client (using pyzeebe to access zeebe). We have to manually restart zeebe and gateway pods.

We hope that if we enable the liveness probe, k8s will restart zeebe gateway and zeebe when the probes fail so we don’t run into the issue we did earlier.

Appreciate any feedback.

Hi @camundaenthu - I thought the liveness probes were enabled by default … I’ll have to double check this. I am not aware of anything you need to be aware when enabling the probes. You can find the documentation here: Gateway health probes | Camunda 8 Docs

Thanks @nathan.loding

This is what I was referring to:

Zeebe section

Zeebe gateway section