Unable to revive connections error

We are experiencing an intermittent error with Kibana which shows as "Unable to revive connection" in our logs. We are running Kibana as a Docker container on the web server. It appears the Cluster continues to take in data properly, but Kibana becomes unusable. Once we bring the container down and back up it starts to work again.

The web server was suppose to be a dedicated coordinating only node. However in production it was accidentally set as "machine learning" and "ingest." Would these two node roles cause any issues with the Kibana docker container?

They might run CPU intensive tasks, so I'd expect performance degradation for the Kibana server. In general, it's a good idea to allocate them on different nodes. However, it doesn't explain network problems. Do you use an orchestrator to shut down and start the Kibana image? Might it affect the network configuration?

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.