Performance Monitoring of Instance and Alerting for CPU & Memory High Utilization

• Navigate to Alarm Creation via Observability & Management menu

Navigation menu > Observability & Management > Alarm Definitions (under Monitoring) > Create alarm

---

• Navigate to Alarm Creation via Compute menu

Navigation menu > Compute > Instances > instance details (eg. bastion-server1) > Monitoring > CPU/Memory Utilization > Actions

Creating Alarm for CPU Utilization (A typical alarm threshold is 90 percent.)

1. Write Alarm name and Alarm summary

2. Select preferred Creation mode currently I go with Basic mode

3. In Metric description select Compute compartment & Interval

4. In Trigger rule select Operator, Value, Trigger delay minutes, Alarm severity, and Alarm body

5. To add additional Trigger rule click on Add aditional trigger rule and when done click Next

6. On the Define alarm notifications page in the Destination section select Destination service, Compartment (where topic is present), and write Notification subject

7. In the Message grouping section if you select Group notifications across metric streams (alarm-level notifications) you will receive only one notification, no matter how many metric streams are firing at a particular time

8. Enable alarm and click Next

9. Review the details and click Create

Creating Alarm for Memory Utilization (A typical alarm threshold is 85 percent.)

1. Write Alarm name and Alarm summary

2. Select preferred Creation mode currently I go with Basic mode

3. In Metric description select Compute compartment & Interval

4. In Trigger rule select Operator, Value, Trigger delay minutes, Alarm severity, and Alarm body

5. To add additional Trigger rule click on Add aditional trigger rule and when done click Next

6. On the Define alarm notifications page in the Destination section select Destination service, Compartment (where topic is present), and write Notification subject

7. In the Message grouping section if you select Group notifications across metric streams (alarm-level notifications) you will receive only one notification, no matter how many metric streams are firing at a particular time

8. Enable alarm and click Next

9. Review the details and click Create

What to do when you receive notication regarding high utilization?

First of all we monitor the metrics and if every thing appears normal then it's okay but if the issue still persists we immediately increase the configuration (Number of OCPUs/Amount of Memory) of instance.

• Navigation

Navigation menu > Compute > Instances > instance details (eg. bastion-server1) > Actions > More actions > Edit

• Scroll down on the Edit instance page where you will find Shape name click Expand button

• Increase Number of OCPUs or Amount of memory as required

• Click Save changes and it will ask you to Reboot instance click on Reboot instance