[ECS] [Autoscaling]: Allow autoscaling rules precise control

### Community Note

* Please vote on this issue by adding a 👍 [reaction](https://blog.github.com/2016-03-10-add-reactions-to-pull-requests-issues-and-comments/) to the original issue to help the community and maintainers prioritize this request
* Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
* If you are interested in working on this issue or have submitted a pull request, please leave a comment

**Tell us about your request**
Allow more precise control over Autoscaling rules for ECS services. Allow to set metric periods and number of datapoints to evaluate an alarm.

**Which service(s) is this request for?**
ECS

**Tell us about the problem you're trying to solve. What are you trying to do, and why is it hard?**
Currently, autoscaling rules for ECS services represented via CloudWatch AlarmLow/High rules for each service, each metric (CPU, RAM, AlbRequests). The problem is that you can not modify these rules to adjust autoscaling the way you want. E.g. period for alarm rule is always 1 minute and 3 out of 3 datapoints for scale out rule and 15 out of 15 for scale in. For our workloads 3 minutes to scale out is an ETERNITY. But description section clearly says "DO NOT EDIT OR DELETE".

**Are you currently working around this issue?**
We tried to experiment long time ago by manually adjusting time periods and number of datapoints in our scaling rules, obviously ignoring a warning in the description and that actually worked, but it is unclear how it will behave in long term usage, so we want an official support for it.

No files, tests, or entry points are named. Start by reviewing the ECS autoscaling request and its current CloudWatch alarm behavior; done means officially supporting configurable alarm periods and datapoints to evaluate for scale-out and scale-in rules.

Written by the indexing model from the issue text.