Add capi support

[QuentinBisson]

This PR adds CAPI support to logging operator

[hervenicol]

You have to differenciate MC from WCs, even on Capi, because the list of logs to retrieve differs.

I was wondering if that could be externalized in a config, so we can select which logs to retrieve for each clusters.
By default, we would have:

• only select logs from kube-system and giantswarm namespaces
• a custom configmap deployed via flux on all MCs (collections?) to retrieve all logs.

Not sure it's part of this PR though 😅

[QuentinBisson]

The différentiation is actually done already in the promtail-config resource as we're using common.IsWorkloadCluster there :)

[QuentinBisson]

@hervenicol this actually looks way better, but that means I have to test it everywhere :D

[QuentinBisson]

Dropping in favor of #81 and giantswarm/object-storage-operator#26