Discrepancy in CPU Utilization Metrics Between ECS Metrics Tab and CloudWatch Metrics for Deployed Microservices

Hi ,
I have deployed my microservices on ECS, and overall, they seem to be working fine. However, I've noticed a discrepancy in the reported metrics between the ECS Metrics tab and the CloudWatch Metrics for one of my services.
In the ECS Metrics tab under services, I am observing the following utilization:

But when I try to check the same service's CPU utilization metrics in CloudWatch, the values appear different:

I'm reaching out for assistance in understanding the reason behind this discrepancy and to determine which set of metrics I should consider as accurate. For instance, I'm observing 96% CPU utilization in the ECS console, whereas CloudWatch is showing only 24%.

Could you please help me understand this difference and guide me on which metrics I should rely on for accurate monitoring of my service's CPU utilization?

No source files, tests, or entry points are mentioned. Start by reviewing the ECS Metrics and CloudWatch views described in the report, then compare their metric definitions and aggregation windows. Done means documenting the reason for the 96% versus 24% readings and which metric should be used.

Written by the indexing model from the issue text.