It should be interessed to extend the monitoring of elasticsearch to monitor trough elastic api:
- Logstash nodes (CPU, memory, uptime, config reloads, events received, events emmited, events latency, ...) - filter for nodes
- Logstash pipelines (no. of pipelines, events emitted rate, no of nodes) - filter for nodes & pipelines
- Kibana (health status, Alerts, memory, max. response time, avg load, uptime) - filter for Kibana instance
Have done some research. Logstash provides a dedicated api for monitoring. There is a full description with curls and examples here: https://www.elastic.co/guide/en/logstash/7.13/node-stats-api.html
This should allow the check if logstash is working very well, uptime, memory and cpu usage and pipelines
For pipeline (https://www.elastic.co/guide/en/logstash/7.13/node-stats-api.html#pipeline-stats) is should be important to know if pipeline is runinng (I think otherwise should not present on output), and the events emitted rate, reload errors, queue type (memory or persited)
But I have to check how to reach the api from elasticsearch server,because the local logstash API will responde only to localhost