Monitor quality

What is checked #

Monitor quality runs a set of checks over your monitors and flags the ones worth tidying. It is a maintenance view, not an alerting view.

• Alert volume — monitors generating far more alerts than the rest.
• Missing recipients — monitors with nowhere to send an alert.
• Missing delay — monitors reading a cloud metric with no delay to let the data arrive first.
• Muted for a long time — monitors muted so long the reason has gone.
• Bad destinations — monitors pointing at a destination that no longer resolves.
• Orphaned composites — composites depending on a monitor that is gone.
• Alerting for a long time — monitors stuck in alert across the week.

The thresholds #

Each check fires at a fixed point, so the same monitors are flagged every time until you change them. The numbers are what make the check worth reading.

• Alert volume — the noisiest tenth of your monitors by alerts sent over the last seven days, and anything above 20 alerts in that period.
• Muted too long — muted for more than 60 days.
• Alerting too long — a group has been in alert for more than 7 days.
• Missing delay — reads a cloud metric with no delay set. A delay of 300 seconds or more is the usual fix, because that is long enough for the metric to land.

Reading the result #

The page reports your total monitors, how many are healthy, and which check each flagged monitor failed. Bulk actions let you mute or unmute a selection in one step.

Working through it #

• Start with bad destinations — those are monitors that cannot reach anyone.
• Then orphaned composites, which are usually left behind by a deleted monitor.
• Then alert volume, muting the noisiest first.
• Leave alerting for a long time alone until the rest is tidied; those are real problems, not configuration.

Onward #

Most of what this page flags is configured in Notification destinations. The monitor types being checked are in Monitors, and the whole category opens at Dashboards and monitors.