Hi There, I an newbie to ELK stack. Appreciate if you can guide/suggest/help about measuring the throughput of ELK cluster in terms of serving the number of messages per second and subsequently autoscaling it in case the throughput is less than the desired-threshold.
Autoscaling is in my view something that should be approached with caution when it comes to Elasticsearch. Whenever you add a node to an Elasticsearch cluster, data will be rebalanced, which adds extra load to the cluster. If you are autoscaling when the load gets high, adding a new node can actually temporarily reduce performance, especially if there is a large amount of data to relocate.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.