[Feature request] Limit throughput when pulling images

### Summary
When the ecs agent is pulling an image that is too big it causes a burst on I/O, on Amazon Linux 2 we see "Stream copy errors" on the logs for docker around the same time big images are pulled.

### Description

Checking /var/log/messages for times when the errors happened

```
grep stream\ copy\ error ../var/log/messages |cut -f 3 -d \ |cut -f 1,2 -d : |uniq -c
2 02:10
2 04:33
2 15:18
2 17:16
2 17:18
18 18:10
58 18:11
6 18:12
2 18:20
16 18:38
2 18:53
```
Checking ecs agent log, I could see some images but one of the images had a size of 3.6GB

### Expected Behavior
I'd like to limit the throughput when pulling images

### Observed Behavior

It pull the image as fast as it can.

Start by reviewing the ECS agent image-pull behavior and the reported Docker stream-copy errors on Amazon Linux 2. Define how throughput should be limited and how large-image pulls should be evaluated against the reported I/O bursts; the issue does not identify implementation files or tests.

Written by the indexing model from the issue text.