[ECS] [feature request]: dynamic resource limits (CPU and memory) of daemon service based on instance type size

### 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**
Option to allocate CPU and memory of a daemon task as a fixed proportion of the total instance CPU and memory.

**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?**
We run daemon service in our ECS clusters, CPU and memory requirement is directly dependent on how big an instance is. The bigger the instance, the higher the number of containers running on it and all of the daemon services workload comes from the number of containers running on it (Eg, logs exporter, metrics exporter, etc).

We set a default CPU and memory in the task definition for each daemon service, the problem we face on the bigger instance is that it either gets OOM or the CPU gets throttled on instances where the number of containers are running more.

**Are you currently working around this issue?**
We provide these tasks enough CPU and Memory so that they can run smoothly on bigger instances. This adds a lots of wastage on the smaller instances where these daemon services runs.

**Additional context**
NA

**Attachments**
NA

The request concerns ECS daemon services whose CPU and memory needs vary with instance size and container count. Start by reviewing the ECS task-definition resource limits and daemon-service behavior described by the issue; done would mean supporting CPU and memory allocation as a fixed proportion of the instance resources without overprovisioning smaller instances.

Written by the indexing model from the issue text.