From blog :
A good rule is to ensure to keep the number of shards per node below 20 per GB heap.
How many shards per node is better to have a low overhead in case of fault ?
This depends a lot on your use case. Can you provide some details?
Logging, log analysis and full text search.
We have 1000 devices more or less that send logs to elasticsearch.
We also use kibana to search details about some events and periodically we create a summary report of these events.
Than I would recommend sticking to the recommendations in the blog post.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.