Hugo Maingonnat

Prefixing the resource type look like the standard way of doing this: `kubectl logs -f deploy/my-deployment` works (but select only one pod). Resources don't always have the `app` label.

Ops, sorry I found out that I need to do it like this : ``` advanced_filter_fields = ['related_obj__id'] ``` It would be nice though it it could work with `related_obj`...

I agree, this was more an example for those looking to add dcgm. However I can try to make it optional, if you think this is possible ? (i am...

It would be a nice feature ! About the RAM tracking, https://github.com/python-recsys/benchy might be a source of inspiration ? It uses memory_profiler package

This probably involve to raise specific exceptions that store the HTTP code instead of raising generic LibCloudError and parsing the code from in the error message https://github.com/apache/libcloud/blob/fe4f81b5a31a8f470ff6ba5af7451cccfc423c2b/libcloud/storage/base.py#L716

It looks like `and getattr(settings, "PYINSTRUMENT_URL_ARGUMENT", "profile") in request.GET` also prevent from using the callback. I feel it should be: ```python if ( show_pyinstrument(request) or getattr(settings, "PYINSTRUMENT_URL_ARGUMENT", "profile") in request.GET...

Looks related to #339 but having a callback would be more powerful

I also have a table that is reaching the `integer` limit on the primary key. Doing the migration to `biging` manually without downtime is quite painful if the table is...

Thank you for the fast answer ! I will try this and let you know if I encounter any issue

If the bucket is dedicated to risingwave it could be simplified to `roles/storage.objectAdmin` (with eventually a condition to `resource.name.startsWith("projects/_/buckets/DATA_BUCKET") || resource.name == "projects/_/buckets/DATA_BUCKET"` ? I found this [documentation](https://github.com/risingwavelabs/risingwave-operator/blob/main/docs/general/state-stores.md), maybe we...