Logstash External Component Retry Attempts

Elastic Stack Logstash
1

Hello Members,

During logstash startup if, for any reason, logstash couldn't connect to the input or output plugin, it keeps on trying indefinitely.

My question is, is there any settings which tells it to stop after specified number of attempts? I'm more interested in Rabbit and JDBC as input and ES as output plugin settings.

UPDATE: For jdbc input, I saw connection_retry_attempts. I didn't see any such settings for rabbitmq input plugin.

Well, Logstash keeps on trying indefinitely. I think, that's OK. But when I send shutdown signal, it's not killing the process. I see SIGTERM received. Shutting down. message in stdout, but it keeps on logging the error message. Ideally, I would expect it to shutdown. Please let me know whether it's a Logstash bug or my understanding is not correct.

Thanks in advance!!

2

The connect! call has a rescue/retry loop that will run forever. The connect! call never ends unless rabbitmq becomes available (or you send a second SIGTERM).

3

Hello @Badger,

Thanks for the response!
I tried sending multiple TERM signals, but still LS kept on logging the connection error message :frowning:

We have a shell script to stop the running process. It keeps on waiting (for some specified time) for process to stop. But for any reason, LS is not stopping, any planned operation after LS stop is getting delayed.
We can always use kill -9, but trying to avoid it as much as possible.