Prometheus metrics

ResticGUI can serve Prometheus metrics at /metrics. It is off by default; turn it on in the configuration:

metrics:
  enabled: true
  # when set, scrapes must send "Authorization: Bearer <token>"
  token: "a long random string"

/metrics is outside the login of the web interface, so set a token whenever the server is reachable by others. A scrape configuration:

scrape_configs:
  - job_name: resticgui
    authorization:
      credentials: a long random string
    static_configs:
      - targets: ["resticgui.example.com:3001"]

A scrape reads ResticGUI's database only; it never runs restic. The repository size and snapshot count are refreshed after each successful backup, cleanup or init.

Metrics

Metric	Labels	Meaning
resticgui_repository_info	repository, backend	Always 1, one per repository
resticgui_repository_size_bytes	repository	Size of the data in the repository
resticgui_repository_snapshots	repository	Number of snapshots
resticgui_job_last_run_timestamp_seconds	repository, kind	When the last job of this kind finished
resticgui_job_last_run_success	repository, kind	1 when it succeeded, 0 otherwise
resticgui_job_last_success_timestamp_seconds	repository, kind	When the last successful job of this kind finished
resticgui_backup_last_duration_seconds	repository	How long the last successful backup took
resticgui_backup_last_added_bytes	repository	Data the last successful backup added
resticgui_backup_last_processed_bytes	repository	Data the last successful backup read
resticgui_backup_last_files	repository	Files in the last successful backup

kind is the job kind, such as backup, forget, prune, check or init; restores are left out. The Go runtime and process metrics are there too.

Alerting

For example, a backup older than a day:

- alert: ResticGUIBackupTooOld
  expr: time() - resticgui_job_last_success_timestamp_seconds{kind="backup"} > 86400
  for: 1h
  annotations:
    summary: "No backup of {{ $labels.repository }} for a day"

Or a last run that failed:

- alert: ResticGUIJobFailed
  expr: resticgui_job_last_run_success == 0

Grafana dashboard

deploy/grafana/resticgui-dashboard.json shows the backups and the repositories: how many repositories there are, which jobs are failing, the age of the oldest last backup, a table with each repository's last backup, size and snapshots, and graphs of time since the last backup, run results, data added, duration, repository size and snapshot count. A Repository filter at the top narrows it to some of them.

To import it, in Grafana open Dashboards, New, Import, upload the file and pick the Prometheus data source that scrapes ResticGUI. To provision it instead, put the file in a folder that a dashboard provider reads:

# /etc/grafana/provisioning/dashboards/resticgui.yaml
apiVersion: 1
providers:
  - name: resticgui
    type: file
    options:
      path: /var/lib/grafana/dashboards/resticgui

It needs Grafana 10 or later.