TheKoguryo's Tech Blog

12.2 Creating Monitoring Alarm

Let’s create an alarm to send an email when CPU utilization reaches a specified number among monitoring set on Compute instance.

• Defined in Observability & Management > Monitoring > Alarm Definition in the navigation menu in the OCI console
• Defined in the options menu of the resource monitoring metric

There are two methods above, but in this document, we will proceed to the second method, which is easy to use because the alarm’s metric is automatically set as the selected monitoring metric.

Create Alarm from Metric

1. Select Resources > Metrics of the Compute instance you set up earlier.

2. In the options of CPU Utilization, click Create an Alarm on the Query

Alarm Definition

Complex alarm conditions are possible with the advanced mode Query, and the following is the basic mode for selecting the UI.

1. Alarm definition

   • Alarm Name: Enter the alarm name to create
   • Alarm severity: select the level, first select the default Critical
   • Alarm body: Alarm message, enter a message to be forwarded by email later

2. Metric description

   • Select Metric name: CpuUtilization.

3. Metric dimensions
   Specifies the scope of the cloud resource to be monitored.
   The Compute instance you selected earlier is set up automatically.
   

4. Trigger rule

   Specify the condition to set the alarm. When the previously set metric criterion (CPU usage average for 1 minute) is 50% or more

Notification Definition

1. Select Notification Target

   Specifies the target topic of the Notification service to be notified when an alarm is triggered. If you already have a topic, you can select it, but if it doesn’t, click Create a topic to create one

   

2. Create a target topic

   

3. Select the topic created as Destination

   

4. Click Save alarm below

Notification - EMail authentication

1. If you set the topic’s subscription protocol to email, a verification email is sent to the target email.

2. Check the received email and click the verification link.

   

3. Subscription confirmation window

   

Notification - Check EMail Topic Status

1. Click the notification target of CPU Alarm created earlier

   

2. Select the topic you created earlier

   

3. Confirm that your email subscription is Active

   

Alarm occurrence test

1. Connect to the Compute instance where monitoring and alarm are set via SSH.

2. Install the stress tool

   sudo yum-config-manager --enable ol7_developer_EPEL
   sudo yum install -y stress
   

3. Perform stress

   sudo stress --cpu N --timeout N
   

Execution example

Adjust according to the number of CPUs of the Compute Instance to be used.

[opc@webserver1 ~]$ sudo stress --cpu 4 --timeout 60
stress: info: [12194] dispatching hogs: 4 cpu, 0 io, 0 vm, 0 hdd
stress: info: [12194] successful run completed in 60s

Alarm occurrence confirmation

1. You can check the history of the created Alarm below.

   

2. Confirm receipt of alarm mail