Python – Programmatically Check AWS EC2 CPU and Memory usage

It's ok not to see that checkbox.

Try restarting the "Amazon SSM Agent" service and then the "Ec2Config" service.

Then, wait to see if your metrics appears. The other config to check is the config.xml in the same folder as the JSON, make sure the CloudWatch plugin is Enabled.

If kjournald is consuming a lot of resource, its mean your OS is doing a lot of journaling operations, which arise from changes to file system (disk operations).

That means something is doing a lot more writing to the file system than it should be, or that there is some issue with your block storage.

Use lsof to see what files are in use by your operating system at any given time.

Failing that, if your Production and Staging environments were created at the same time, with the same instance class, in the same AZ, and you're confident that your PHP application is well designed (should be OK if its Drupal) and you have not changed anything recently, I'd consider re-provisioning the EC2 instances.

Make a snapshot and re-deploy, ideally on a different instance class to ensure you're moving to a different infrastructure stack.

I presume the instances are EBS backed, in which case they could be using some block storage in AWS that has gone bad.