Anna Kapuścińska

The first two points were implemented in #1797, there are two other items left. (In the initial issue I wrote `ServiceAccount` instead of `ServiceMonitor`, sorry :woman_facepalming: that was a mistake.)...

> > Don't start the operator metrics server if tetragonOperator.prometheus.enabled = false > > @lambdanis not really get the point on this? should be a new flag that we can...

Hey, sorry for the delayed response > `tetragonOperator.prometheus.enabled = false` with the logic in args I believed operator metrics server doesn't start right? > https://github.com/cilium/tetragon/blob/5f2247c4b677aa3202ea0d0081a02da0f0e75c95/install/kubernetes/tetragon/templates/operator_deployment.yaml#L36-L41 > if yes, this issue...

@yukinakanaka Thank you! Sounds good to me, feel free to open a PR. Since the operator doesn't make sense outside of k8s context, it's fine to disable the metrics server...

`tetragon_data_events_total` and `tetragon_data_event_size` metrics were eventually added in #1224 and refactored later. They seem to behave, so I'm closing this PR. Thanks for opening it @swastik959.

It would be very useful to add a PR template indeed. I think it can be inspired by the [Cilium PR template](https://github.com/cilium/cilium/blob/main/.github/pull_request_template.md), maybe a bit simpler. Things that should be...

I'm not super convinced by the `_stats` convention, I feel something like `_meta` would be more intuitive. Just a comment, it's not super critical :) > Rename all bpf map...

The bugs here are fixed, so closing the issue. Further improvements to how Tetragon monitors BPF maps (including monitoring more of them) should be done as part of https://github.com/cilium/tetragon/issues/2778

Tetragon operator changed a lot since this issue was opened - now it's deployed as a separate deployment instead of an init container, and it has resources configurable via Helm...

#1495 is now merged, so this PR can be rebased to iterate over what was added there.