k8s operator service improvements

[pmbrull]

1. The omjob operator is not cleaning up pods after the "ttlSecondsAfterFinished" is over. It runs indefinitely as per the event log

 Normal ExitHandlerCreated 46s   omjob-operator Created exit handler pod: om-job-123b421c-357b-49a2-9e8d-7fd6dfa23a2f-5decd8bf-exit
 Normal ExitHandlerCreated 36s   omjob-operator Created exit handler pod: om-job-123b421c-357b-49a2-9e8d-7fd6dfa23a2f-5decd8bf-exit
 Normal ExitHandlerCreated 26s   omjob-operator Created exit handler pod: om-job-123b421c-357b-49a2-9e8d-7fd6dfa23a2f-5decd8bf-exit
 Normal ExitHandlerCreated 16s   omjob-operator Created exit handler pod: om-job-123b421c-357b-49a2-9e8d-7fd6dfa23a2f-5decd8bf-exit
 Normal ExitHandlerCreated 6s    omjob-operator Created exit handler pod: om-job-123b421c-357b-49a2-9e8d-7fd6dfa23a2f-5decd8bf-exit

This happens regardless of the main pods lifecycle (completion or failure). This creates a huge chunk of completed or Error pods in our cluster waiting to be deleted.

2. Lets add the ability to handle tolerations on ingestion pod creation