It should be possible for IRecordProcessors to mark themselves for shutdown. about amazon-kinesis-client

[coldear]

This is exactly what I am looking for.
Ideally, when user request to destroy current processor, KCL should be able to create a new one for the same shards and start to process from the last checkpoint.

[vlanka]

Thanks for your feedback. We will look into adding this support in the future.

[ridha]

+1

[rhodey]

you can accomplish this pretty easily with Guava's ListenableFuture or Java8's CompletableFuture & the factory pattern & I'm happier with this being app-specific logic instead of something forced by the library-- amz has a history of doing threading & executors weird :/

you already have to implement IRecordProcessorFactory so give yourself a base abstract pattern to start, like my KinesisRecordConsumerFactory

my KinesisRecordConsumer has some application specific stuff in there but you should be able to get the idea.

then in your main class add a callback to KinesisRecordConsumerFactory.getErrorFuture() & kill your app or shutdown your worker or whatever you want.

[jqr]

It would be great for the MLD to have this as well. We're currently calling exit(1) which works, but also causes the whole MLD to die, taking all owned shards with it. Restarting with workerId means these come back up quickly, but it's overkill.

[pfifer]

Thanks for reporting this. We're investigating this feature, and will prioritize with other feature requests.

Others - Please comment or add your reaction to this post so our team is aware of how to prioritize this change.

[BobbyJohansen]

I would agree with this seems like an important piece of functionality to provide to the implementer

[zeynepsu]

We're tracking this internally. Closing this issue.