Description of problem:

The customer encounter an issue trying to trigger an upgrade to 4.18.
The clusterversion was able to create the job related to start the update to the new version, but the related POD was never created (job status remained empty)

After the excessive jobs from the ETCD (+90K objects) have been removed, the cluster was updated without any issue.

Version-Release number of selected component (if applicable):

Issue was faced when trying to update from 4.17 to 4.18

How reproducible:

Still working on a reproducer.

Steps to Reproduce:

1.
2.
3.

Actual results:

The update to the next minor version is not triggered

h2 .Expected results:
The update job should not be stuck waiting for the POD to be created.

Additional info: