Submit a ticketCall us
Home > Success Center > Network Performance Monitor (NPM) > Alert timings

Alert timings

Table of contents
Created by Richard Casey, last modified by MindTouch on Jun 23, 2016

Views: 31 Votes: 1 Revisions: 3

Overview

This article provides information regarding alert timings.

The alerts in Orion operate based on certain timings both in the alert itself and the polling frequency.

Environment

NPM11.5 and later

Detail

The alerts are evaluated from the polled (database data).
The node down (Status) alert is triggered by a loss in ICMP packets. By default we poll the nodes (ping) every 120 seconds and if a poll is missed then the device is set into a warning state and we then fast poll the node for a further 120 seconds and if still no response then the node is set to down so this process can take up to 4 minutes and then any delays in the alert set up would be added to this value.
The interface status is obtained using SNMP polling of the node and they would be marked as down if the node returns this status after an SNMP result, so if an interface goes down then it may not be 'seen' for up to 9 minutes with the default polling.
In the new alerting each alert is evaluated every 15 seconds (configurable) so the database status of the relevant resource is checked every 15 seconds.
The traps are sent to Orion and are not polled so would be more real-time than the polled data. It is possible to trigger an alert based on matching trap data using the Trap rules in the Trap viewer.

 

Last modified
21:45, 22 Jun 2016

Tags

Classifications

Public