DocumentaciónHelp

Escalation and automatic closing

Who ARGUNIS notifies when nobody takes an incident, when it closes on its own, and when it stays open until someone closes it.

On this page

This page explains what ARGUNIS does with an incident while nobody is handling it, and how it ends.

Escalation: notifications when nobody takes it

When an incident opens, the notification goes to the assigned resolver group (or, if the group has on-call, only to the person on call). If nobody acknowledges or takes it, ARGUNIS notifies more people according to the group's escalation policy. The minutes are counted from when the incident opens, or from when it is reassigned.

The default policy is:

Severity

First step

Second step

Critical

Group leads at 15 min

Company admins at 60 min

Warnings

Group leads at 1 h (60 min)

Company admins at 2 h (120 min)

Each group can have its own policy, with up to 5 steps per severity. See Configure a group's escalation.

  • Each step is sent only once, at its time, even if no new events arrive.

  • Each step is recorded in the incident's notes.

  • If the group has no leads, the leads step goes to the admins.

  • During a maintenance window, escalation is postponed until the window ends.

What stops escalation

  • Acknowledging or taking the incident.

  • Snoozing it (for as long as the snooze lasts).

  • The incident being resolved, on its own or manually.

When you reassign, escalation starts counting from zero again with the new group's policy. When a snooze period expires and the problem has not ended, the incident goes back to Open and escalation also starts from zero.

When it closes on its own

When an incident closes on its own, ARGUNIS sends the recovery notification to everyone who received the opening or an escalation.

Incident type

How it closes on its own

Status: service down, server without data, CPU, RAM, disk, UX test, database, probe checks, and similar

When its check is OK again (it recovers)

Log errors, agent findings, and probe findings

After a period without new events (30 min by default; adjustable between 5 min and 24 h). Log errors also close together with the recovery of the service of the same component, once there are no new errors

Security

After 60 min without activity

APM latency and service without data

When they return to normal

APM errors (new errors and error rate above threshold)

After a new deployment of the service in which the error does not reappear within 30 min, or after the APM thresholds' days without recurrence (14 by default). For the error rate, the service must also be below the threshold

An administrator adjusts the time without new events in Closing event incidents.

When it stays open

An incident stays open (and keeps escalating, if nobody took it) as long as the problem continues. In particular:

  • APM errors are not closed by the log errors' time without events, nor because the error rate drops below the threshold. They stay open until someone closes them with Declare a fix or with Resolve, until a new deployment leaves them without the error, or until the days without recurrence pass (14 by default). When closed, the error becomes Resolved in the APM: if it appears again, a new reappearance incident opens. See APM alerts and incidents.

  • A status incident stays open while its check keeps failing, even if no new events arrive.

Related

Search Argunis Monitoring

Type at least 2 characters to search Argunis Monitoring.

↑↓ to navigate · Enter to open · Esc to close