Notifications¶
ARGUS posts to a webhook when something happens that a person should know about. Configure it under Settings → Notifications.
Configuring the webhook¶
Set the webhook URL and save. Leaving the field blank keeps the current webhook unchanged; to remove one, use the explicit clear option rather than blanking the field.
Point it at whatever your team already watches — a chat integration, an alert router, or your own receiver.
Routing to more than one place¶
Settings → Notification routing sends notifications to different webhooks by cluster group, minimum severity and what happened — for example high and above from production clusters to an on-call channel, everything else to a team channel.
| Part | What it is |
|---|---|
| Destinations | A name and a webhook URL. The URL is shown only while you enter it; afterwards ARGUS reports only that one is set, because a webhook URL usually carries its own access token |
| Rules | Considered in order. Each matches a cluster group, a severity floor — High and above includes critical — and a kind of event, and names a destination |
| Default rule | Always last and always matches, sending to the organisation webhook above. Nothing is ever dropped |
Saving replaces the whole table in one step, so notifications are never routed by half an edit. Off by default: with no rules, everything goes to the organisation webhook.
What it is for¶
Notifications are for things that stop progressing without a person:
- a high-severity incident opening
- a plan waiting for approval, and one that has been waiting too long
- an execution that failed
- a pull request that has sat unmerged
- a validation that could not confirm an outcome, or an incident reopening
A plan that has been in the queue for days is escalated rather than silently waiting, so an approval queue nobody is watching becomes visible instead of becoming a backlog.
Keep it narrow¶
Route notifications narrowly at first. A quiet, trusted channel is more useful than a busy one people learn to ignore, and ARGUS is capable of being noisy on a fleet that has not had its detector settings tuned.
If notifications are too frequent, the fix is usually upstream: tune detectors, add tolerations for accepted conditions, or set storm protection. See What ARGUS detects.