Steps to reproduce

1. Create a package with a known good state.
2. Wait for it to go to Ready.
3. Modify the package to a known bad state (ex: use a uri in the sso client id)
4. Observe that the package gets stuck in Pending without doing additional retries.

Expected result

Package should retry up to 5 times and then either get marked as Ready or Failed.

Actual Result

Package is stuck in Pending until a restart of the watcher pod.

Severity/Priority

Medium - this is very confusing to end users on what is happening.

Additional Context

Our shouldSkip logic filters out pending packages. During a retry we leave the status in Pending, resulting in it getting stuck unless it hits other cases (max retries or first time the watcher has seen it).