Simon Pasquier

hmm I'm still getting the error with the SHARD env var being unset. Output of the `kubectl describe pod ...` command: ```` Last State: Terminated Reason: Error Message: level=info ts=2024-07-25T12:41:43.520299622Z...

https://github.com/thollander/actions-comment-pull-request?tab=readme-ov-file#permissions

Hmm it looks like there's no easy solution :disappointed: https://stackoverflow.com/questions/61419552/workaround-to-post-comments-from-github-actions-from-forked-repos

maybe https://docs.github.com/en/actions/writing-workflows/choosing-when-your-workflow-runs/events-that-trigger-workflows#workflow_run could work but it would still be convoluted to pass the PR id to the second workflow.

I don't think that `pull_request_target` will work because it's only reacting on events like "PR opened", "Review requested", ... but nothing like "CI has ended with status=X".

I'm definitely not expert in dual-stack networking. IIUC status.podIP resolves to the IPv4 address of the pod while the endpoint address resolves to IPv6? The alertmanager-operated service is a bit...

> I usually expect that a pod listens to either :: (or 0.0.0.0 in an IPv4-only environment), or on localhost if they need to be locked down to just the...

The `--cluster.listen-address` doesn't look correct to me. Trying something similar on my local env, I see errors in the Alertmanager logs: ``` ts=2024-06-27T13:20:09.864Z caller=cluster.go:179 level=warn component=cluster err="couldn't deduce an advertise...

@m0untains can you share the Alertmanager logs and see if you have the same error as reported [here](https://github.com/prometheus-operator/prometheus-operator/issues/6611#issuecomment-2194680221)?

Thanks for confirming. Using `status.podIPs` instead of `status.podIP` fixes the issue "by accident" but it's not a proper resolution for the project. Unfortunately there's no "easy" way to retrieve the...