Announcing support for adding multiple alarm severity levels in one ...

We're happy to announce an enhancement to alarm definition capabilities for the Oracle Cloud Infrastructure (OCI) Monitoring service. You can now add multiple severities in one alarm for the same metric.

Multiple alarms on the same metric

We understand that you often need to define multiple severities within the same alarm, driven by the criticality of the threshold value that gets breached. For example, you might want to create an alarm to get Warning notification with CPU utilization greater than 90%. However, if it goes beyond 95% you might want to receive Critical notification for the same alarm. Previously, you needed to set up two separate alarm rules, one for the Warning alarm and the second for Critical alarm.

Now, with the ability to create up to two severities in one alarm, you can conveniently and efficiently manage resources with fewer alarms, reducing the notifications noise. It also means that we're expanding the alarm feature to support multiple trigger rules. Each alarm condition is represented by a trigger rule, having its own set of attributes, such as monitoring query languages (MQLs), severity, alarm body, and pending duration.

Each condition acts a child alarm, run by priority. An alarm is in OK state only when the values for all the metric streams are below the threshold for all the alarm conditions. We trigger the alarm notifications and maintain history when all the conditions in the alarm have been evaluated.

Trigger rules

You can decide the processing order of each condition by the order you define those severities, so that when the condition with higher precedence is met, the alarm notifications for condition of lower precedence don't triggered. You can see the trigger rule based on multiple queries defined in MQL in various flows-Alarm list, Alarm details, Alarms history-and while viewing the metrics stream status.

Notifications

As part of the enhancements, each MQL has a separate Alarm body field. You now have the option to send a separate message for each condition. When the "Group notification" option is enabled, a notification is sent when the highest priority condition is met. A notification isn't sent even if other resources are breaching the lower priority conditions. However, with the "Split notifications" option enabled, a separate notification is sent for a resource in a Critical state and other resources in Warning states.

Alarm history

Currently, history is maintained at the alarm level, such as the alarm status changes from OK to FIRING or FIRING to OK, in the context of one metric condition and one severity. With the new changes, we can now provide a reference to the condition that triggered the alarm to switch states.

Getting started

The enhanced ability to add multiple severities in one alarm is available in all commercial regions. These features don't cause breaking changes to your existing alarm definitions, even with one condition and severity. For more details, refer to the technical documentation at multiple severity alarms.

We welcome you to sign up for the Oracle Cloud Free Trial or sign in to your account to experience Oracle Cloud Infrastructure Monitoring for monitoring your infrastructure. If you have any questions or feedback, contact us. We'd love to hear from you.

Public link

Please use the above public link if you want to share this noodl on another website.