Adding ECK Operator Liveness Probe

[MattThomson]

I've attempted to add liveness and readiness probes to the ECK operator itself (not the Elasticsearch,kibana,etc pods), but I can't find any documentation on exposed endpoints for health checks. So far I'm doing this with a kustomization patch:

- op: add
  path: /spec/template/spec/containers/0/livenessProbe
  value:
    initialDelaySeconds: 10
    failureThreshold: 6
    tcpSocket:
      port: 9443

However, while this seems to work, it results in the pod logs reporting

2021/11/02 14:39:01 http: TLS handshake error from 10.240.0.14:50396: EOF

Every time each probe runs (default 10 second intervals). Is there another way to set up liveness/readiness probes for ECK, or are they not recommended for ECK?
e.g. an endpoint which can be used with httpGet instead of tcpSocket? I've tried the same port with "/", but this returns 404.

[michael.morello]

Hi,

It's a good question. I'll open an issue in the ECK repo to discuss it (I'll update this thread with a link to the issue)
I think there are a few points to discuss, using the webhook server might not be good idea as I'm not sure it reflects the status of the operator.

Thanks

[michael.morello]

I created this issue to discuss it further.