Set up alerts & notifications
Dockstash makes failure loud. Any failed backup, restore, or drill triggers an immediate alert email to the project owner, and a dead-man’s-switch heartbeat catches the silent case where nothing errored because nothing ran. Every alert carries the project, the error, the last-good timestamp, and a deep link.
How to do it
Know what fires an alert
Four kinds: a failed backup (any layer — dump, restic, disk space), a failed restore, a failed restore drill (byte or row mismatch), and a heartbeat alert when a project has had no successful backup inside the heartbeat window. You do not configure per-failure rules — failures always alert.
Tune the heartbeat window
The heartbeat threshold (default 24 hours) is the "silence is failure" window: if a project’s last successful backup is older, you get an alert — including projects that have never succeeded. Alerts dedupe within the window and reset the moment a backup succeeds.
Add a webhook for your on-call flow
Set an alert webhook URL and every alert also POSTs its JSON payload there — plug it into Slack, PagerDuty, or your own bot. Email and webhook carry the same payload, so automations see exactly what you see.
Use the audit trail
Destructive and billable actions (backup runs, restore triggers, schedule deletes, plan changes) append to an immutable audit log with actor, action, target, IP, and timestamp — there is no update or delete path. Review it from your account or via the audit API.
Do it in one click with Dockstash
Dockstash runs the exact dump above, restics it off-site, and drill-tests the restore automatically — no script to maintain.
Last updated: July 2026
Related guides
Troubleshoot failing backups
Read live restic logs, let stale locks fix themselves, revive offline agents, and understand heartbeat alerts.
Run restore drills
Automatically restore the latest snapshot into a scratch workspace, byte- and row-diff it against the source, and get a pass/fail badge.
Redis
Back up Redis in Docker
MySQL
Back up MySQL in Docker
Frequently asked questions
Will I get spammed if a project keeps failing?
Heartbeat alerts dedupe for the length of the window before re-alerting, and a successful backup clears the dedupe immediately. Failure alerts fire per failed run — if runs keep failing, that is signal, not spam.
What is in the alert payload?
The project name, the error detail, the last-good timestamp, and a deep link to the project dashboard. The webhook receives the same fields as JSON.
Why did I get an alert when nothing failed?
That is the heartbeat: no successful backup landed inside the window. A paused schedule, a stopped agent, or a project that never ran all look identical to silence — the dead-man’s switch exists exactly for those cases.
Can external monitoring watch Dockstash itself?
Yes — point your uptime monitor at the unauthenticated health endpoint (200 when the stack is healthy, 503 when degraded). Aggregate counters are exposed on a metrics endpoint with no PII.