Alerts and on-call
Copies this page as Markdown with context, so an AI can walk you through your case step by step.
Choose where you want to hear about problems and what happens if nobody answers.
Channels
| Channel | What you need | Plan |
|---|---|---|
| An address. Your account email is added automatically | All | |
| Discord | A channel webhook: Edit channel → Integrations → Webhooks → New webhook | Starter+ |
| Telegram | Send /start to @ganyant_bot and paste the chat ID it replies with. In groups, add the bot and type /id (group IDs start with -) | Starter+ |
| Slack | An Incoming Webhook from the Slack app | Pro+ |
| Coming soon | Pro+ | |
| Webhook | An HTTPS URL of yours. See Signed webhooks | Pro+ |
| SMS | Coming soon | Agency |
| PagerDuty | An Events API v2 integration key. Outages resolve automatically on recovery | Agency |
Use Test next to each channel to send a test alert.
What gets notified
- A monitor going down and recovering.
- SSL certificate about to expire (30, 14, 7 and 1 days).
- New high or critical security findings.
- Disk over the configured threshold on a server.
- SSH or RDP brute force and new listening ports.
- Failed backups.
On-call rotation (Pro+)
Define an ordered list of people, each with a channel and an escalation time. When an incident opens, every active channel is notified. If nobody clicks Acknowledge within the configured time, the alert escalates to the next person on the list, and so on.
Retries
If a channel fails (for example, a webhook that is down), we retry with exponential backoff up to 6 times over about 30 minutes.
If alerts don’t reach you
- Use Test on the channel: if the test does not arrive, the problem is the channel; if it does, check that the monitor is not paused.
- Email: check your spam folder and add
alertas@ganyant.comto your contacts. - Telegram: if you blocked the bot or removed it from the group, send
/startagain and update the chat ID. - Discord and Slack: if the webhook was deleted on the server, create a new one and replace the URL.
- Outage alerts wait for confirmation from several regions; a one-off failure from a single region does not notify.