Week one with the app: every notification feels important. Week four: your phone buzzes so often that you mute it during lunch — and miss the one alert that was a genuine emergency. The problem is not the app. It is the absence of a policy.
Alert fatigue is a documented failure mode of every monitoring tool, and a phone in your pocket makes it personal. The fix is a deliberate split: a tiny set of events that may interrupt you anywhere, and everything else reviewed on a schedule. This page proposes that split for an Endpoint Central fleet, and the review habits that make it work.
An interruption is for events where delay makes damage. A management server that has stopped talking to the fleet. A critical, actively dangerous vulnerability appearing across many machines — the kind you would approve out of cycle. A device reported stolen, where the lock and wipe actions in the app are time-sensitive. That list should stay short enough to recite. Everything else — and it is a long, legitimate list — is information, and information belongs to a review habit, not to your lock screen. Write the list down where your team can see it; an alert policy that lives in one admin's head changes every time that admin is on leave.
The app makes scheduled review easy because the data is glanceable: system health status, counts of vulnerable computers, missing patches, automated deployment progress. Build two fixed moments into the day — a morning check before the office fills up, and a late check before you leave — and let everything non-critical wait for them. Patch approvals are the classic example: approving in two calm batches beats approving in fifteen buzzes, and you will read what you are approving more carefully. Batching also creates a rhythm users learn: IT visibly reviews the fleet twice a day, every day.
Your first policy will be wrong in both directions — something noisy will sneak into the interrupt bucket, and something important will be missing from it. That is normal; schedule the correction. After a month, ask three questions. Which alerts did I act on immediately — do they deserve their place? Which did I swipe away unread — move them to the review bucket without guilt. And which problem reached me by a user's phone call instead of by the app — that one earns a promotion. Two or three tuning rounds produce a phone that buzzes rarely and means it every time, which is the only kind of alerting that survives year two.
Watch the app surface fleet health and pending approvals in the full video — the raw material your alert policy organises.
▶️ Watch Full VideoNo. Missing patches are a standing condition of any real fleet, reviewed in your scheduled checks. Reserve interruptions for the rare patch situation where waiting until the next check creates real risk.
The one that tells you the management system itself is failing — server down or agents unable to reach it. A monitoring tool that cannot report its own death will let every other problem grow in silence.
It is safer, provided you agree who acts. Duplicate awareness with a clear owner for each incident beats a single phone that was on silent, in a drawer, or on leave.