Create an alarm policy
Make an alarm open an incident only if it crosses a threshold or lasts a few minutes, stay as an event, or be ignored.
Use a policy to cut down the noise. For example: "CPU on the servers in the Payments group opens a critical incident only if it stays above 90 for 10 minutes; otherwise, it stays as an event."
Before you start
You need Administer on Core (Administrator role or a company role with that level).
Steps
Go to Core → Alarm policies.
Click New policy.
In Condition (supports *), type the condition or pick one of the common ones below the field. Examples:
cpu,mem,disk*,latency,ssl,down,errors,sec:*,vuln:*.In Scope (empty = whole company), narrow it down if you want:
Specific asset.
Inventory group (includes subgroups).
Source type.
Priority (tiebreaker): among equally specific policies, the highest one wins.
In What to do, choose Incident, Event only, or Ignore.
If you chose Incident, fill in these optional fields:
Threshold: the alarm value must be equal or higher (applies to server CPU, RAM, and disk). Below the threshold, it stays as an event only.
Duration in minutes (1 to 1,440): the alarm must last that long; if it recovers sooner, it stays as an event only.
Incident severity: The alarm's, Critical, or Warning.
If you want, write a Note explaining why the policy exists.
Leave Policy enabled checked and click Create.
The policy applies to new alarms; anything already open runs its course. Every change is recorded in the audit log.
Edit, pause, or delete
Click Edit on the row, change what you need (uncheck Policy enabled to pause it), and click Save.
To delete it, click the trash can icon.
What happens next
In Core → Events, alarms that stayed as event only or that are waiting out the duration show a tag with the rule number.
On the incident, the Alarm policy field shows which rule opened it.