A question that we answer quite often is: What's the best way to monitor key performance metrics in Elasticsearch—such as response time? There are several open source projects for #Elasticsearch monitoring tools, and one very good commercial solution. In this article, we invite you to take three minutes our of your day to acquaint yourself with the best monitoring tools.
At Qbox, we manage many thousands of clusters—for thousands of customers. And, we enable seamless integration of many excellent open source tools. When we rank according to the choices of our customers, these are the most popular (in order of usage):
- Roy Russo's Elasticsearch HQ (royrusso/elasticsearch-HQ)
- Elasticsearch Kopf (lmenezes/elasticsearch-kopf)
- Big Desk (bigdesk for elasticsearch)
These are quite notable, but less popular:
- Elasticsearch Whatson (xyu/elasticsearch-whatson)
- Elastic Hammer (andrewvc/elastic-hammer)
- Elasticsearch Inquisitor (polyfractal/elasticsearch-inquisitor)
- Elasticsearch Paramedic (karmi/elasticsearch-paramedic)
In addition, we can't avoid mentioning the tool built by Elastic, the fantastic Marvel. This is a commercial add-on, and they license it in blocks of 5 nodes. At Qbox, several customers use Marvel with their Qbox endpoints, though it is a BYOL model.
All of these tools are quite useful for preemptive diagnostics on performance and scaling issues, but they don't provide continuous monitoring and data persistence. Marvel also shares this limitation with the other monitoring plugins: an engineer or tech must monitor the metrics manually to get event notifications. To supplement these tools, Qbox continuously checks clusters for potential problems, and will notify users when any problems arise. Read more in our article, Automatic Cluster Alerting on Qbox.