Tune down unnecessary producer exception

[PeterN]

Is your feature request related to a problem? Please describe.
If a producer is idle longer than ~8_000_000ms - possibly only in azure environment -, it logs a "Kafka Producer Exception". which can incorrectly lead to an assumption that the connection is broken for good and it won't be publishing the next time it needs to.
However, this is not true as automatic reconnection happens underneath.

Describe the solution you'd like
Tune down the exception, so it is only shown when relevant, such as failure on next publish.

Describe alternatives you've considered
Keep alive, but excessive and isn't a guarantee either. Only been able to reproduce using either producer/kafka in Azure cloud.

Additional context