Metrics for scaling elasticsearch nodes horizontally or vertically

We are using Elasticsearch cluster of 3 data nodes 3 master nodes 1 client node. configuration 4GB RAM and 1.5 TB disk for each node and replication is 1 and 6 partition for data nodes.

We currently delete indices which are not neccessary after 7 days , which might change in future(meaning retention period might increase). After deletion 350GB of data is present always in cluster, which will grow eventually. Prior to deletion around 700GB of data is present.

There are few metrics like tripped metrics, pending tasks which tell the state of cluster if it needs to scale.
Is there a list of metrics available which can give foresight of mode additional nodes if needed in future.

Usually GC and heap use is the one to keep track of, if you find nodes are pushing >80% heap use they will GC a lot.

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