The alert you silenced in February.
The scenario:
A disk-space warning fired daily for a week in February. It was noisy, low-risk, and mid-crisis, so someone muted it. It has been silent ever since. So has the disk it was watching.
Muted alerts are decisions with no expiry date. The stack quietly accumulates them until the one alert that mattered turns out to have been silenced eighteen months ago by someone who left.
The suppression list is also an inheritance problem. Alerts muted by people who have left carry no context, so nobody dares unmute them and nobody owns them. The estate changes underneath the suppression, and the rule safely noisy in February is silently masking a real failure by August.
The prompt:
You are building an alert hygiene review for an MSP monitoring stack.
Context: [your monitoring tools, roughly how many alert rules exist, when suppressions were last reviewed]
Build:
The suppression audit: how to export every muted, snoozed, or downgraded alert across the stack, with owner and date where recoverable
A triage rule for each suppression: reinstate, tune the threshold, or retire the rule, never leave as-is
The expiry standard going forward: every suppression gets an owner and a review date at creation
A monthly 20-minute review slot, and the one metric: suppressions older than 90 days, target zero
Silence is not health. Prove the difference this month.