Private Cluster unable to start (Timeout on startup)

[jan-schumacher]

Describe the bug
When a Cluster is not able to connect to the stable (googleapis) repository the helm-operator container is not able to start up because the request is hanging in a 30sec timeout and the readinessprobe is not getting ready because the healthz endpoint is not reachable. Also with disabled repositories this is still the case.

To Reproduce
Steps to reproduce the behaviour:
0. A K8s Cluster

1. Cut the internet connection or add a firewall rule to disable connection to googleapis.com
2. install helm-operator and fluxcd

Expected behavior
The readinessprobe gets ready although there is no connection possible. A configurability for the probes would be handy or a real disable option for repositories

[hiddeco]

This is due to the hardcoded import I added in #124. The workaround till the permanent solution is in place is to mount an empty file at the path it imports (/var/fluxd/helm/repository/repositories.yaml).

[hiddeco]

This has been fixed in #141.