Monitors and alerts

Get an email or webhook when a service's CPU, memory, disk or error rate stays above a threshold.

Add a monitor

Open the service's Metrics tab and select Monitors, or go to Settings > Monitors.
Pick a metric: cpu (% of the limit or of the host), memory (% of the limit), disk (volume % of the host filesystem, or GB), or http_5xx_rate (% of requests).
Set the threshold and how many minutes it must hold (1 to 60).
Pick channels: email to chosen members and any project webhook.

Monitors are checked every minute. A monitor fires once when it enters the alarm state (monitor.triggered) and once on recovery (monitor.resolved). The service card shows a red dot while it fires.

New Postgres, MySQL, Redis and MongoDB services get a disk monitor at 80%.

Next steps

Email and notifications