Vigil

Alerts & quiet hours

Alerts are sent through channels — create them once under Alert channels, then attach any combination to each monitor.

Channel types

  • Email — plain-text alerts to any address.
  • Slack / Discord / Microsoft Teams — via each platform's incoming webhook URL; setup hints are shown in the form.
  • ntfy — push to any ntfy topic (ntfy.sh or self-hosted). Protected topics can authenticate with an access token or a username/password (self-hosted). Down alerts are sent at urgent priority.
  • Pushover — needs your user key and an application API token. Down alerts are sent at high priority.
  • iOS push — native push notifications to every iPhone signed in to the team through the Vigil app. The option appears once at least one device has signed in.
  • Generic webhook — POSTs incident JSON to your own endpoint, optionally signed with HMAC-SHA256. See Webhook payloads.

Every channel has a Send test button that fires a sample alert so you can verify the destination before you need it.

What you get notified about

  • Downtime — always on. Includes cause, HTTP status and which locations confirmed it.
  • Recovery — on by default; includes total downtime duration.
  • SSL expiry — once per certificate when it enters the warning window.
  • Domain expiry — once per expiry date before the registration of an HTTP or DNS monitor's domain runs out, and immediately if the registry puts the domain on hold, into redemption or up for deletion.
  • DNS record changes — when a DNS monitor's records change but still pass. On by default; a change that fails is downtime and alerts as such.
  • Degraded — a third-party provider reporting degraded performance, or a DNS monitor's nameservers misbehaving while the answers still pass. Optional (on by default for DNS monitors).
  • Bot-protection challenges and redirect changes — optional warnings, off by default.

Quiet hours

Under Settings → Quiet hours, owners can define a daily window during which team channels stay silent. Incidents are still detected, recorded and visible on timelines (with a "suppressed" note) — you just aren't pinged. Status-page subscribers are not affected by quiet hours: they are your audience, not your on-call rotation.

No false alarms

Alerts only fire for confirmed incidents: repeated failures per location, agreed by the location quorum you set per monitor. Maintenance mode silences a monitor entirely while you deploy. See Monitors & checks.