We are using logstash with kafka + elasticsearch and running multiple pipelines which works great!
However, under certain circumstances kafka input will receive an exception and the pipeline will terminate.
If all pipelines are kafka input, all pipelines will terminate which also causes logstash to terminate and it will be restarted by systemd which is not perfect but it is ok. However, when not all input plugins are kafka input, logstash will not stop and terminated pipelines will not restart.
Is there anyway to automatically recover from this by restarting single pipelines? Or perhaps a hack to fix it?
Kind regards /Johan