Closed w32-blaster closed 4 months ago
is there any reason why this PR isn't reviewed at all by the maintainers?
@w32-blaster The integration tests are failing because the way that the tests are currently written, no execution_context
is supplied. If you want me to fix up the tests in this commit, let me know.
Please do it. I don’t have enough context to fix that. Thanks
@robbavey, can you help me to fix this please?
is it necissry that client_id unique across multiple piplines? our logstash cluter distribute on 3 server and use the same client, it seems nothing wrong.
The current implementation makes the client_id unique when using multiple threads, but when using one more more threads with multiple pipeline configurations, we are getting the error
javax.management.InstanceAlreadyExistsException: kafka.consumer:type=app-info,id=logstash-0
This is because every pipeline is trying to use the same client_id to connect to kafka. A workaround is to explicitly specify the client_id using a unique string.The patch bellow is adding an additional suffix to the client_id after the thread # which is the pipeline_id. With this, we can be sure that we don't use the same client_id in a multi pipeline setup.