Logstash multi-pipeline output failure breaks other pipelines

Elastic Stack Logstash
1

Running Logstash 7.8.1

Multi pipelines

Pipeline A output to B and C
B output to ES cluster 1
C output to ES cluster 2

Shutdown ES cluster 1
Error from output attempt to send bulk request... normal error when ES is down

However, I would expect A -> C to still flow, but it does not, the entire flow is now stalled.

Is this expected? I thought the pipelines worked independently?

Thanks.

2

No, that is the whole point of the output isolator pattern. Are you using persistent queues? The in-memory queues may back up enough to cause the pipeline to block.

3

Yes - we use in memory queues. Persistent queues don’t scale well when spread across a large cluster dropping files to disk everywhere.

I feel like this is a bug?