Alert fatigue
Design alert rules people will trust — not mute.
Start with one high-signal rule
Ship a single ERROR spike or threshold rule before adding heartbeat and pattern matchers. Validate with a sample log from Integration.
Use cooldowns
Set cooldownMinutes so repeated failures from the same deploy do not open duplicate incidents.
Maintenance during deploys
Snooze rules or create a maintenance window when shipping migrations. Document the window in your status page if customers depend on uptime.
Channel routing
- Slack for actionable ERROR spikes
- Email for daily digests only
- On-call rotations on Scale when coverage matters