[FLINK-39970] Retry incomplete JobManager deployment deletion

[stellagai]

What is the purpose of the change

Kubernetes deployment deletion waits can fail or time out before the old JobManager deployment is fully removed. The operator currently logs these failures and continues reconciliation, which can submit a replacement cluster while the old deployment is still terminating and result in AlreadyExists errors.

Brief change log

• Propagate non-404 errors while waiting for Kubernetes resources to be deleted.
• Retry reconciliation instead of creating a replacement cluster before deletion completes.
• Preserve the best-effort JobManager shutdown behavior before mandatory deployment deletion.
• Update deletion error and timeout tests.

Verifying this change

This change added and updated tests and was verified with:

JAVA_HOME=/opt/homebrew/opt/openjdk@17/libexec/openjdk.jdk/Contents/Home \
  mvn -pl flink-kubernetes-operator -am \
  -DskipITs \
  -Dtest=AbstractFlinkServiceTest \
  -Dsurefire.failIfNoSpecifiedTests=false test

Tests run: 40, failures: 0, errors: 0, skipped: 0.

Does this pull request potentially affect one of the following parts:

• Dependencies (does it add or upgrade a dependency): no
• The public API, i.e., is any changes to the CustomResourceDescriptors: no
• Core observer or reconciler logic that is regularly executed: yes

Documentation

• Does this pull request introduce a new feature? no
• If yes, how is the feature documented? not applicable

[Dennis-Mircea]

Thanks for the PR. The fix itself is correct, but this overlaps heavily with #1138 (FLINK-39953), which targets the same root cause in the same deleteBlocking wait catch and the same shutdownJobManagersBlocking caller.