FAQ
Are monitors shared across projects?
No. Monitors, incidents, alarms, channels and playbooks are all scoped to a single project. Switch projects from the console's project picker (or the portal header) to manage a different environment.
What counts as "Down"?
A check fails when it doesn't meet its type's success condition — a bad status code, a connection/timeout, a DNS record that doesn't resolve to the expected values, or a certificate past its expiry threshold. A check that exceeds its Timeout also counts as Down. Use an alarm's Consecutive failures to avoid alerting on a single blip.
Why didn't I get a notification when a monitor went down?
Notifications only fire when an alarm on that monitor opens an incident, and only to the channels the alarm targets. Check that: the monitor has an alarm, the alarm is Active, it lists at least one channel, and that channel has at least one subscriber. Send a test notification from the channel page to confirm delivery.
What's the difference between an alarm and an incident?
An alarm is the rule (the condition + severity + channels). An incident is a single event it opens when the condition is met — the thing you acknowledge and resolve.
Do playbooks run automatically?
Only when they are armed and attached to an alarm that fires. New playbooks are created disarmed and never run until you arm them from the detail page — see Playbooks.
It says "Beta" — is it safe to use?
Yes, but Healthcheck & Alerting is still in beta: features and limits may change and you may hit the occasional rough edge. Avoid arming destructive remediation playbooks against production traffic until you've validated your alarms.