Look into handling unavailable kafka service/ broker

[mehab]

If for some reason, the topics that the application wants to send data to are temporarily unavailable then the application should be able to hold on to the events for a certain time and retry like a polling mechanism until the topics are available again. This would help in stopping any events being lost due to unavailability of service.

[nscuro]

I think this should be covered by the Kafka clients automatically, see https://www.conduktor.io/kafka/kafka-producer-retries and #70 (section "Producers").

Granted, we'll probably need to tweak and verify the configuration.

[nscuro]

ToDo: Provide some sane defaults (e.g. reducing retry attempts to something less than MAX).

[nscuro]

Not pursuing this anymore as a consequence of #1619