Simon Pasquier

And to be clear, we have a consensus that it would be useful but someone needs to work on it (see #5485).

The cause of the issue is that the addition or update of a secret/configmap only triggers a reconciliation of Prometheus/PrometheusAgent objects living in the same namespace as the secret. But...

> I wonder if we want to work on a separate branch to make reviews easier and only merge to main only when the whole feature is ready 🤔 I'm...

> The goal is to break this PR into smaller ones to make it easier to review, and not to release it experimentally. It could be both ;) We don't...

Before jumping ahead of ourselves, I've added the discussion to the next office hours.

> I also have a doubt, should we include a changelog entry even if it's behind feature-gate and only half implemented? It's your call but if it's not ready for...

> > > While reviewing, I wonder what should happen for Prometheus rules. Should they be deployed to scaled-down shards? My intuition says no (for instance, alerting rules usin `absent()`...

> I discussed with Josh and we agreed that an appropriate fix would be instead of calculating if the alert was firing or not by checking [Resolved()](https://github.com/prometheus/common/blob/v0.42.0/model/alert.go#L78) indirectly in template.NewData(),...

One downside would be the additional gossip traffic on every group interval? But I imagine that Alertmanager could refresh the notification log when it approaches the (last update + data...

Looking at 66a0ed21bdb0720b4ba083d35acd6ae77fa7b0b5 , stringSlice was added exactly for this use case. > Does it make sense for Remove to require []string instead of ...string IIUC Remove is expected to...