Inter-Node Response Times

Is there a metric I can track which will tell me how long nodes are taking to respond to pings and cluster state updates?


No, this is not something that Elasticsearch tracks. You could work it out with sufficiently verbose logging. Why do you want to know these metrics? I'm not sure they tell you very much about the overall performance of the system.

Am seeing periodic very high response times so want to track those response time to see the behaviour in more detail.

What exactly are you measuring here? Do you mean searches and/or indexing are slow? Pings and cluster state updates are both interactions between each node and the elected master, and this is normally not something that happens during searches or indexing, so it's unlikely that these metrics will help with your investigations.

I'm seeing periodic very high latency (around a minute) on pings and cluster state updates. This is a new environment so isn't very loaded

Ok, turning your original question round, how are you seeing the high latency on these messages? Logs? If so, can you share the messages you're seeing?

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