Pergunta

I want to deploy tools to monitor my Production Percona servers (5.6 on Linux). I never used Nagios or any tool before. I did go through this great link but I wanted to clarify few things before deploying them:

  1. Those mysql exporter and node exporters of Prometheus which I need to deploy in my database host - what would be overhead of them? Resource intensive?
  2. I did go through Nagios as well - Is Nagios better than Grafana and is it possible to use Grafana with Nagios?

Any better suggestions would really help me as well.

Foi útil?

Solução

  1. The 'mysql' & 'node' exporters are lightweight processes and shouldn't impact your servers if you're running with default settings. Be careful when enabling the additional stats which use information_schema - these can impact your DB servers (especially when you have a lot of databases / tables)

  2. Nagios is chiefly used for 'Alerting' while Grafana on the other hand is chiefly used for 'Trending'. The alerts that Nagios provides are based on specific thresholds (warning / critical). It is possible to use Prometheus' AlertManager (which collects data for Grafana) in order to receive alerts on specific metrics. With Prometheus you can leverage the time series database to generate more intelligent alerts which look at the rate of change rather than a specific threshold - see https://prometheus.io/docs/alerting/alertmanager/ for more details on how to set this up.

Licenciado em: CC-BY-SA com atribuição
Não afiliado a dba.stackexchange
scroll top