ability to skip messages that are failing/crashing

[vigith]

For serving cases, there is no point working on a message that is crashing the UDF causing restarts. This will decrease the availability if we get stuck on a failing message. We should skip those messages. The only way to do that is by looking at the redeliver counter generated by the ISB.

NOTES

1. This is a message loss scenario
2. An offending message might end up droping more than one message (if batch-size > 1)