Drain tasks when a thread errors

[docktermj]

Is your feature request related to a problem? Please describe.
A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]

When SENZING_EXIT_ON_THREAD_TERMINATION is true

https://github.com/Senzing/redoer/blob/6d193de879247a8782f73deafe7f705024b66137/redoer.py#L138-L142

The code immediately kills the program.

https://github.com/Senzing/redoer/blob/6d193de879247a8782f73deafe7f705024b66137/redoer.py#L1653-L1657

The problem with this approach is that it could kill "in-flight" messages. Under normal circumstances this is OK because the "in-flight" messages would not be committed to Kafka and a different consumer would be given the message after a period of time.

However, if the code is changed to send a commit to Kafka before guaranteed delivery to Senzing, then killing the program could kill "in-flight" messages for each thread.

To minimize this, when a thread terminates and SENZING_EXIT_ON_THREAD_TERMINATION is true, then allow each thread to complete (including the monitoring thread) and then exist the program.