Avoid unlimited defaults for max_concurrent_producer_ids & max_transactions_per_coordinator

[jason-da-redpanda]

These parameters are meant to be safeguards for excessive memory usage but we default them to unlimited.

In particular have seen memory allocation errors due to max_concurrent_producer_ids being unlimited (including ..the infamous : Seastar - Failed to allocate 1081336 bytes )

These are the defaults for 23.3.10 ...

rpk cluster config get max_concurrent_producer_ids 
9223372036854775807
rpk cluster config get max_transactions_per_coordinator
9223372036854775807

We should consider setting defaults to a sensible value ..

JIRA Link: CORE-2020