Exiting a Kafka group occurs without service failure

[Nekit2217]

Greetings. Kafka periodically throws silent errors that block the service and prevent it from crashing. For example

Logs

2024-10-07T13:34:04 [INFO] - kdi-req-events: Seeking consumer to partition offsets: [1:20614721801,0:20633486622,2:20548970596,]
2024-10-07T13:35:28 [WARN] - kdi-req-events: LockClientError::VersionAlreadyExists(2691227)
2024-10-07T13:35:28 [WARN] - kdi-req-events: LockClientError::VersionAlreadyExists(2691228)
2024-10-07T13:35:28 [WARN] - kdi-req-events: LockClientError::VersionAlreadyExists(2691229)
2024-10-07T13:35:28 [WARN] - kdi-req-events: LockClientError::VersionAlreadyExists(2691230)
2024-10-07T13:36:16 [INFO] - kdi-req-events: Delta version 2691231 completed in 54284 milliseconds.
2024-10-07T13:36:55 [INFO] - kdi-req-events: PRE_REBALANCE - Revoke
2024-10-07T13:39:35 [INFO] - kdi-req-events: POST_REBALANCE - Assign [1, 2, 0]
2024-10-07T13:39:35 [INFO] - kdi-req-events: Handling rebalance assign. New assigned partitions are [1, 2, 0]
2024-10-07T13:39:47 [INFO] - kdi-req-events: Seeking consumer to partition offsets: [1:20614773023,0:20633535515,2:20549022982,]
2024-10-07T13:46:21 [INFO] - kdi-req-events: Delta version 2691355 completed in 57398 milliseconds.
2024-10-07T13:55:41 [INFO] - kdi-req-events: lazy_load_identity;
%4|1728311124.136|MAXPOLL|rdkafka#consumer-1| [thrd:main]: Application maximum poll interval (1800000ms) exceeded by 285ms (adjust max.poll.interval.ms for long-running message processing): leaving group

After this, the service stops ingesting and simply stands idle

[rtyler]

I believe I know what this is but it sure is hard to reproduce. I think we need a panic handler that will allow a thread to panic and crash the whole process when that happens.