Alert Performance
Alerts are "active" server components causing constant CPU load. The CPU load caused by an alert is a multiplication of several factors:
- Number of alert's triggers.
- Number of contexts checked by each trigger. For example, in a large installation a trigger targeted to
users.*.devices.*context mask will cause much higher load that a trigger targeted tousers.john.devices.dev1. - The CPU load impact of alert trigger's expression.
Once an alert is raised by one of its triggers, an additional performance impact can be caused by its corrective actions.
Alerts don't cause significant constant memory usage. However, considerable spot memory usage can be caused by evaluation of trigger expression referring a large dataset or execution of a corrective action.
Was this page helpful?