How to Monitor CPU Credits on AWS T3 Instances Before Throttling Hits?

You can monitor T3 CPU credits using CloudWatch metrics. It's recommended that you set up proactive alerts so that you have some time to make adjustments to your environment before all of the CPU credits are exhausted.

These are the main CloudWatch metrics that you can monitor to keep tabs on your CPU credit usage, CPUCredit Balance and CPUCreditUsage.

CPUCredit Balance will show you the remaining CPU credits that you have, and CPUCreditUsage keeps track of credit consumption per minute.

It's best to set graduated alerts, so getting an alert at 30%, and then followed by another alert at 10%. This will let you know if the issue is temporary, or if you have a process that's consuming more credits that it should be. You can create a visualization on a CloudWatch dashboard so that your operations teams can track this as a part of your standard monitoring of your AWS environment.

Thanks for the question, and feel free to reach back out!