An ElasticSearch availability and performance monitoring plugin for Nagios.

This plugin works by submitting API requests to a local or remote ElasticSearch server. ElasticSearch servers will respond to these API requests by default. If yours don't, check that you have not set http.enabled to false.

This plugin may be used to monitor ElasticSearch data and proxy nodes.
It is important that you run one instance of this plugin for each data node in your ElasticSearch cluster. Certain failure modes that can be localised to a particular data node will not be reported by unaffected data or proxy nodes.

ElasticSearch defines its own thresholds for 'green', 'yellow', and 'red'. Most other ElasticSearch monitoring plugins simply take this 'health colour' and map it directly to a Nagios check status (OK, WARNING, or CRITICAL, respectively). While that approach will work (insofar as it will report a problem that ElasticSearch has detected), an operator is going to need more information to narrow a real problem down to its root cause. ('Why is my cluster yellow?')

This monitoring plugin requests detailed operational data from ElasticSearch and uses that information to derive its own 'health colour'. If our health colour concurs with ElasticSearch's health colour, then we can be reasonably sure we know what is wrong with the cluster and may report that problem to the operator.