Walther Lee

Sorry, I mentioned it in a thread in the kube-state-metrics Slack channel and forgot to put it here. The bug isn't exactly in metrics_handler if `Run` runs only once, which...

That's a good point, I can look into that. I think `m.Run` could probably be just `m.ConfigureSharding`, but discovery doesn't have access to the Statefulset or k8s client to calculate...

I've made some changes to use `m.ConfigureSharding` in discovery and to run `m.Run` only once. I'm testing it on a test cluster and so far it's working fine. I'll let...

@CatherineF-dev I added new changes to run `m.Run` only once when the pod starts, and to reconfigure sharding in discovery instead of recreating the IndexInformer. I also removed some of...

A summary of changes per file to help with the review: - `internal/discovery/discovery.go`: Use m.BuildWriters(ctx) instead of m.Run(olderContext) to rebuild writers instead of recreating the whole handler. BuildWriters cancels the...

I can confirm that this is still in v0.29. Here's a little more context. The problem is that the alertmanager resets the timer before flushing the alerts ([here](https://github.com/prometheus/alertmanager/blob/f7b4feffd3c8d4570e7daed234c4ba56b9ef8b05/dispatch/dispatch.go#L569-L574)), even if...

@midnightconman I'm not sure I follow what kind of problem this is trying to solve. Labels and annotations have to come from Prometheus, so the webhook integration has to handle...

@djmcgreal-cc Yeah, I'm actually wrong, the alertmanager already supports templating in some URL fields, for example, [here](https://github.com/prometheus/alertmanager/blob/v0.29.0/config/notifiers.go#L406). I think adding support for webhook makes sense too. Labels and annotations should...

Sure, I can look into that. I'm pretty sure they don't overwrite, so that works too. On the other hand, without a flag, it will be a breaking change for...

@primeroz has the issue occurred again? Even in v0.20.0 the silence should stop the notification pipeline if all alerts are muted