The secondawaitTermination call aftershutdownNow() can throwInterruptedException, but this exception is not caught here. If interrupted, the exception will propagate to the outer catch block at line 1691, which will callshutdownNow() again (redundant since it was already called) and set the interrupt flag. However, the success log at line 1686 would never be reached if interrupted.

Consider wrapping this secondawaitTermination call in its own try-catch block to handle interruption properly, or document that interruption during forced shutdown is acceptable.

[nitpick] The timeout has been increased from 5 seconds to 60 seconds per executor shutdown. SinceshutdownScheduledExecutorService is called three times during broker shutdown (lines 1508, 1603, 1604), this could potentially extend total shutdown time from 15 seconds to 180 seconds (3 minutes) in the worst-case scenario where all three executors time out.

While this provides more time for graceful task completion, consider:

1. Making the timeout configurable viaBrokerConfig to allow operators to tune it based on their workload
2. Adding metrics or logging to track actual shutdown durations to identify if such long timeouts are necessary
3. Documenting the maximum expected shutdown time for operators

Note: The test atBrokerShutdownTest.testBrokerGracefulShutdown currently expects shutdown within 40 seconds, which would need to be updated if this long timeout is retained.