Operate · Alarms & incidents

An alarm counts against a datacenter only when it belongs there.

Every alarm is scoped: inside the site, on an edge link leaving it, on something it depends on, or downstream. Downstream symptoms never count on their own; they show under the incident that caused them. Each incident arrives with affected services and customers, ranked root-cause hypotheses and actions.

Who opens it

From a thousand alarms to one incident.

Scope decides whether an alarm counts against a datacenter at all; the incident carries the cause, the impact and who is handling it.

NOC engineer asksIs this our fault or a symptom of someone else's?

The scope ring: site, edge, dependency or downstream. Downstream symptoms sit under the incident that caused them.

Incident manager asksWhat changed before it broke?

Ranked root-cause hypotheses with their evidence, and the change cited by the leading one.

On-call lead asksWho is on call, and when is the next customer update due?

The rota, the paging chain, the next escalation and the customer notice that is due.

On the screen

What the screen holds.

The scope rings around a datacenter beside one incident with its cause, its impact and the symptoms downstream.

Also on the screen

On call and communications

Rota per region, the paging chain, the next escalation, and customer notices with their update cadence.

Problems

Incidents grouped by a shared cause, such as one optics lot behind several link failures.

See your own estate in it.

Bring the fault your team keeps dealing with. We walk it end to end on a demo estate shaped like yours, then agree what a one-site pilot should prove.