Setup a prometheus box with alertmanager

Setup a box with prometheus and alertmanager ( requires mail setup ). For collecting our monitoring data from gitlab and other sources:

Systemd ( packaged )

• failed systemd unit
• CPU resource accounting research

https://github.com/prometheus/node_exporter ( not sure if it does cpu accounting)

Prometheus-postgresql-exporter

• package exporter
• add ansible role for enabling exporter when postgresql_servers
• setup alerts for postgresql

Prometheus-mysql-exporter

• package exporter
• add ansible role for enabling exporter when mysql_servers
• setup alerts for mysql
• setup grafana dashboard percona dashboards

Nginx

• 200,400,500's etc. (parses logs)

Use loki instead, with alerting

Memcached - packaged

• Cache/hit ratio's etc

prometheus-memcached-exporter

General

CPU/Load status

https://github.com/prometheus/node_exporter

UWSGI

Not monitored but status are exported, needs packaging

Btrfs

Btrfs errors can be monitored using a custom textcollector