I've noticed that when I scale-out the number of Logstash apps I have in containers writing to same ES cluster it stops all activity (index & search)
But as soon as I go back to the reduced number of workers (from 20 to 8) the cluster works back again. The problem is that # of apps is not enough to cope with the load and I want to send more data to ES.
why is this happening ? I though it was because I didn't have enough ingest nodes (my Logstash apps are writing to 2 of them) but then I no longer think that because it is "bulk" thread pools in many data nodes (not ingest) who are completely loaded and rejecting requests.
appreciate any inputs,