Skip to content

Notifications

Stay on top of what your platform is doing without watching the screen. Every user configures their own channels from the bell menu in the header (Notifications modal).

What triggers a notification

Job outcomes — a deploy, rebuild, backup or any other operation finishing (successfully or not) — and infrastructure alerts (host unreachable, disk critical, instance down…). Failed deploys and rebuilds are treated as severe and always delivered immediately.

Your preferences

  • Level — All (every job outcome), Failures only (default), or None.
  • Muted projects — silence everything from projects you don't care about.
  • Notifications only cover what you can already see: project visibility rules apply to every channel.

Channels

In-app (always on)

Toasts for finished jobs and critical alerts, plus the alert badge in the header. Real-time — no configuration needed.

Email

Toggle it on and pick a mode:

  • Immediate — one email per job outcome.
  • Daily digest — a single summary email every morning. Severe failures (deploy/rebuild) still arrive immediately.

Telegram

  1. Create a bot with @BotFather and paste the bot token.
  2. Send any message to your bot, then click Detect chat ID — the platform finds your chat automatically.
  3. Send test to verify. Done: job outcomes arrive as Telegram messages.

Webhook

Point the platform at any HTTPS endpoint to integrate with Slack bridges, on-call tools or your own automation. Two event types arrive as JSON POSTs — job_state_change for successful job completions, and alert for failures and infrastructure alerts:

{
  "event": "job_state_change",
  "job_id": 123,
  "job_name": "Deploy Instance",
  "state": "done",
  "severe": false,
  "host": "host-1",
  "instance": "production",
  "log_url": "https://…/cloud/log/123",
  "timestamp": "2026-07-12 10:00:00"
}

{
  "event": "alert",
  "alert_id": 45,
  "code": "job_failed",
  "level": "critical",
  "message": "Deploy Instance on production failed",
  "host": "host-1",
  "instance": "production",
  "job_id": 123,
  "log_url": "https://…/cloud/log/123",
  "timestamp": "2026-07-12 10:00:00"
}

When an alert clears automatically (the monitor that raised it observes recovery), the same payload shape is sent again with "event": "alert_resolved" — consumers can close the incident they opened on the alert event. Dismissing an alert by hand in the panel does not send this event: the operator silencing it is already looking at it.

If you set a secret, every request carries an X-IncubaCloud-Signature: sha256=<hex> header — the HMAC-SHA256 of the raw body. Verify it with a constant-time comparison before trusting the payload.

Backup usage alerts

Separately from job notifications, each backup backend can email a warning when bucket usage crosses a threshold (default 80%). Configure it on the backend detail page.