Skip to main content

Monitor & alert

Two halves of the same job. Monitoring gives you metrics and logs on Grafana dashboards for when you go looking; alert rules tell you by email or Telegram when something happens so you do not have to, and both work the same here as on every other engine FPT Database Engine manages.

info

The full procedures are documented once, for every engine:

warning

Not available on a Sharded Cluster. Its detail page has no Monitoring tab — the tab set is Overview, Backup, Resource, Notification, and Action logs. Notification and alert delivery still work; the Grafana dashboards do not.

Monitoring​

  • Monitoring is not self-service. It must be activated for your account by FPT Support. Until it is, the Monitor tab shows a screen asking you to contact the support team, so raise the request before you need the data rather than during an incident.
  • Integrate a workspace to start collecting metrics and logs; Disintegrate to stop. The Integrated Status column reflects which state each workspace is in.
  • Metrics and logs are read in Grafana, reached from the dashboard URL on the workspace detail page. Logs use Loki as the data source.

Alerts​

  • Alerts cover backup and restore results, resource thresholds, and maintenance notices.
  • For a Resource alert you set vCPU, RAM, and storage thresholds. An alert fires when any one is met, so meeting all three sends three notifications.
  • Configuring a recipient is a prerequisite, not a nicety: the backup service refuses to activate until one exists.
warning

Telegram does not support the Maintenance service. Maintenance notices, including the 7-day and 1-day warnings before a maintenance window, can only reach you by email. Configure at least one email recipient even if you prefer Telegram otherwise.

Next steps​