Cluster monitoring

@tvc_apisani The link to Elasticsearch exporter is in GO and it’s available as docker.

Point 2) do you think that will be easy to migrate from ELK+XPack to Opendistro in the future?
Elastic already prevent migrate Kibana to OD for 7.10.2 due to flattened mapping usage.
In the near future, it’s easy to implement a “change” in index mapping and prevent the shift from Elastic Elasticsearch and OpenDistro.

The link to Elasticsearch exporter is in GO and it’s available as docker
I know. That’s the reason I was argumenting it could be a bit complicated for some users to setup such configuration: they would need ODFE + docker exporter + prometheus (standalone or via docker) + grafana (standalone or via docker) : not exactly a 1-2-3 setup compared to xpath (i repeat: for some users). As i said above: it is mainly a matter of which kind of users ODFE would primarely target: the average user of the tech-savv user (or, to simply put it: users with small clusters and relative experience or IT staff of large deployments).

Elastic already prevent migrate Kibana to OD for 7.10.2 due to flattened mapping usage.
Really ? Didn’t known about this one. Interesting.

Going through the thread I have a question regarding Prometheus Exporters:
If, for example in this proposal, I would like to add a new stats endpoint for a new OpenSearch metric, this would require a new exporter for OpenSearch. Does anyone see value in forking the exporter in favor of the OpenSearch project for additional reasons? If so - where would you suggest the exporter be maintained?

1 Like

From my perspective any software can export prometheus metrics to be scraped as needed. Jaeger does this for example and it allows for the easy collection of data about the operation of the system. This allows for agentless collection of the metrics in a standard format: Monitoring Jaeger — Jaeger documentation (jaegertracing.io) There are also standard ways to do this with mixins which is another way to create a nice monitoring system for OpenSearch : 2018-02 Prometheus Monitoring Mixins Design Document - Google Docs

1 Like