Skip to main content

Outage Alerts

Your Kuploy instance continuously verifies its own health — including its control-plane database, backup archiving, and backup freshness — and reports to kuploy.app while everything passes. If those reports stop for more than about 15 minutes, kuploy.app emails your team automatically.

Because the alert is triggered by silence, it works even when the outage is total: a crashed platform, a dead cluster, or a broken network all raise the same alarm. The email is sent from kuploy.app's own infrastructure, so it does not depend on anything running inside your cluster — including your own mail server.

Who receives alerts​

Alert emails go to everyone responsible for the instance, across both your kuploy.app account and the instance itself:

  • Your kuploy.app account: the tenant owner and any account members with the admin role.
  • Your instance's Platform Admins: everyone listed under Admin → Platform Admins with the Owner or Operator role.

The two lists are combined and de-duplicated; each person receives their own copy.

Designating who gets paged​

To add someone to the alert list, add them as a Platform Admin on your instance (Admin → Platform Admins → Add Platform Admin) with either role — Operator is appropriate for on-call staff who should respond to outages but don't need platform configuration access. Changes take effect within about an hour.

To remove someone, remove their Platform Admin entry (or their kuploy.app admin membership).

What an alert looks like​

  • Subject: [kuploy] <your instance>: "pg-email-health" has gone silent
  • When: within roughly 15 minutes of your instance going quiet.
  • One alert per outage — you will not be flooded; when the instance recovers, reporting resumes automatically and the next outage alerts again.

Health is reported per category — Email, Platform, and Infrastructure each have their own check — so the alert subject tells you which part of the platform stopped reporting, not just that something did. The same categories appear as status strips on your instance's license page at kuploy.app, alongside a 30-day uptime view; see Continuous Backup & Point-in-Time Recovery for what each strip shows.

Coverage starts at birth: your instance registers its checks with kuploy.app automatically on its first sync, so a brand-new deployment whose reporting never starts raises an alert just like an established one that stopped.

If you receive one, treat it as real: the instance either failed its own health checks, is down, or is unreachable — all worth immediate attention.