Read the Monitoring page, filter by severity, and tell whether the alert ingest is actually running.

1. 1

   Open Monitoring

   Choose Monitoring in the sidebar. The five cards across the top are Alerts Today, Critical, Errors, Warnings and Unreviewed. Each card is clickable and sets the severity filter beneath it.

   

2. 2

   Filter to what you are working

   Severity pills are Critical, Error, Warning, Info and OK. Critical, Error and Warning are on by default; Info and OK are off, because they are noise on a working shift. The Status pills are All, Unreviewed and Reviewed. Tick Include decommissioned only when you are investigating history — it is off by default so retired sites do not clutter the feed.

3. 3

   Let the feed poll

   The feed checks for new alerts every 30 seconds on its own. The status line under the AMMP Alert Feed heading tells you where it is: Ready, Checking for new alerts…, or Last checked with a time. New alerts are only prepended while you are on page 1 — on any other page the counter moves but the list does not, so you do not lose your place mid-read.

4. 4

   Turn sound on if you are on watch

   Sound On / Sound Off persists between sessions. It only fires for batches containing Critical (a double beep) or Error (a single ping). Warning, Info and OK are deliberately silent, so the sound means something when you hear it.

5. 5

   Check the ingest is alive

   The Last AMMP ingest chip in the feed header is your health check for the alert cron. If it reads hours ago rather than minutes, alerts have stopped arriving and the problem is the ingest, not a quiet night. Pull from AMMP forces a manual pull for every linked site and reports how many were new.