Ayoub Mrini

I addressed your comments @bboreham, if you could take a look.

the added `TestDelayedCompactionDoesNotBlockUnrelatedOps` should fail due to https://github.com/prometheus/prometheus/pull/13754 I'll adjust that.

I added a `waitingForCompactionDelay` check there as well. cc @bboreham I think this is ready for review now.

I ran into this as well while running `promtool tsdb dump` prometheus: 2.46.0, promtool: 2.46.0 Here is the panic: ``` 2023-09-14T12:00:21.636554378Z panic: expected newly cut file to have sequence:offset 2:8,...

This is easy to reproduce: Just run a Prometheus instance with: ``` global: scrape_interval: 1s scrape_configs: - job_name: prometheus static_configs: - targets: ['localhost:9090'] ``` And a dump in a loop:...

Should we close this as fixed by https://github.com/prometheus/prometheus/pull/11642?

Thanks for your response. I think the tests you added are sufficient to validate this. I think we can close this as fixed by https://github.com/prometheus/prometheus/pull/11642 @roidelapluie

I think we could make use of informers' `WatchErrorHandler` https://github.com/kubernetes/kubernetes/blob/862ff187baad9373d59d19e5d736dcda1e25e90d/staging/src/k8s.io/client-go/tools/cache/shared_informer.go#L185-L198 to surface some errors via a metric. I'll give it a try.

[partialMetadataRemoveAll](https://github.com/search?q=repo%3Acert-manager%2Fcert-manager+partialMetadataRemoveAll&type=code) doesn't seem to handle special `cache.DeletedFinalStateUnknown` objects as expected, see: https://pkg.go.dev/k8s.io/[email protected]/tools/cache#TransformFunc for details. This should be the same issue as the one fixed by https://github.com/prometheus-operator/prometheus-operator/pull/6298 The fix is straightforward...

/cc @simonpasquier as we've already talked about this.