Definition
An alert is a notification triggered by a monitoring system when a condition is met — most often when a check fails (downtime) or succeeds again (recovery). It's how monitoring gets a problem in front of a human quickly.
Alerts can go to many channels: chat tools like Telegram, webhooks into other systems, email, or SMS. A good alert is timely, actionable, and routed to where the responsible person will actually see it.
Why It Matters
Detection is useless if nobody hears about it. Alerts turn a recorded failure into immediate action, shrinking the time between an outage starting and someone fixing it. Well-tuned alerts are the difference between catching a problem in minutes and discovering it hours later from customers.
How It Works
You configure which conditions trigger an alert (a failed check, repeated failures, a slow response) and where it goes. When the condition is met, the monitor sends the notification through the configured channels. Recovery alerts fire when the condition clears, closing the loop so responders know it's resolved.
Real-World Example
A site's API goes down. Within a minute the monitor sends a Telegram alert to the team's channel and fires a webhook into their incident tool. The on-call engineer sees it instantly, fixes the deploy, and gets a recovery alert confirming the API is back.
Best Practices
- Route alerts to channels your team actively watches
- Alert on confirmed failures to avoid noise from blips
- Include enough context to act (what failed, when, where)
- Send recovery alerts so responders know when it's resolved
- Tune thresholds so alerts stay meaningful and trusted
Common Mistakes
- Sending alerts to a channel nobody monitors
- Alerting on every transient blip until people tune them out
- Omitting context, so responders waste time investigating
- Not sending recovery notifications
- Having no alerts at all and learning of outages from customers
In Monitoristic
Monitoristic sends alerts via Telegram and webhooks the moment a check fails, and a recovery alert when the endpoint is back. Route them into your team's chat or into automation tools so the right person acts immediately.