Report throttled time for task level cpu quotas

### Summary
CFS throttling metrics are not available

### Description
Using task level CPU quota creates a cgroup with CFS quota as a parent cgroup of docker container. We would like to see how often our tasks gets throttled by the scheduler to take appropriate actions and tune our services quotas accordingly. However ECS agent does not report this metric and the one that docker itself provide also not helpful as it does not know about the parent cgroup apparently.

### Expected Behavior
We should be able to access `ecs.service.cpu.throttled_time` metric in cloud watch.

### Observed Behavior
Our tasks gets throttled and not metrics get reported.

Start by reviewing the ECS agent behavior described in the issue, including the parent cgroup created for task-level CPU quotas and the existing Docker metrics. Determine how a CFS throttled-time value could reach CloudWatch as ecs.service.cpu.throttled_time. The payload names no repository files, tests, or implementation entry points, so completion criteria require confirming the metric is reported.

Written by the indexing model from the issue text.