Fix upgrade operator helm chart
When upgrading Helm could produce following errors:
Error: UPGRADE FAILED: rendered manifests contain a new resource that already exists. Unable to continue with update: existing resource conflict: namespace: stackgres, name: stackgres-operator, existing_kind: /v1, Kind=Secret, new_kind: /v1, Kind=Secret
Now, this happens because the pattern used in the operator is that of creating resources using helm depending on some values or to create them using kubectl from a job. Now when upgrading if we switch from kubectl created to helm created due to changes in those values (like for example providing a certificate). This will not work making the upgrade fail with the error above.
Helm is able to patch resources created with helm so we have to create resources with helm and patch them with the initialization / upgrade jobs in order to enable helm upgrades. There are some exceptions in the case of those resources that are never created by helm (like csr or jobs that are binded to a pre/post /install/upgrade hook).