Alerts
When Rippler tells you, who it tells, and how to make it tell you less.
When an alert fires
Two things are worth telling you about, and Rippler sends one notification for each:
- Down — the monitor stopped reporting. Either grace ran out on an expected
ping, or the job pinged
/fail/. - Up — it recovered from being down.
A recovery is only sent for an outage that was actually announced. If a monitor went down and came back without ever reaching your failure tolerance, you get neither message, rather than an "all clear" for an alarm that never rang.
Nothing is sent for a monitor in grace. That is the point of grace: a ping running
a little late is not news.
Where alerts go
To every member of the team that owns the monitor. A monitor belongs to a team rather than to a person, so there is no single owner left to send it to.
Each person controls their own delivery with the alert emails setting on their profile. Switching it off stops that person receiving alerts; it does not change what the team alerts on, and it does not affect anyone else. The team members list shows who has it off, so an admin cannot quietly turn a team's alerting into nothing.
Discord
Set a Discord webhook URL on a project and its monitors post there too, as an embed carrying the monitor name, its project, its schedule and a link back to it. This is per project, not per team, so a noisy project can go to its own channel.
Both channels are attempted for every alert, and each delivery is retried a few times with backoff, so a blip at Resend or Discord does not silently lose a notification.
Making alerts quieter
The wrong fix for a noisy monitor is to pause it — that stops the monitoring too. The right one depends on why it is noisy:
| Symptom | Fix |
|---|---|
| The job is fine but finishes later than expected | Raise the grace duration to a little above its worst normal run time. |
| The job genuinely fails sometimes and recovers on its next run | Raise failure tolerance, so it takes several failures in a row to alert. |
| You do not want this person emailed | Turn off alert emails on that person's profile. |
| The job is not supposed to run for a while | Pause the monitor. No pings expected, no alerts, and paused time is not counted against its uptime. |
Raising failure tolerance only delays the notification. The monitor still shows as down in the app the whole time, so nothing is hidden from someone looking. A success resets the count, so two failures a week apart never add up to one alert.
Missed runs count as failures too: with a tolerance of 3, a daily job that stops running entirely alerts on the third missed day.