[ECS] [request]: Preserve Instance Store Data on ECS Managed Instances

### 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**
What do you want us to build?
This issue proposes a capability for ECS Managed Instances that preserves instance store data.

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

**Background**
Customers running AI inference workloads on ECS cache large model weights on local NVMe instance store for low-latency serving. Each cache rebuild can cause wastage in idle GPU compute and may cause prolonged service degradation. This forces customers to choose between managed infrastructure benefits and local cache preservation.

**Potential Feature Proposal
I am opening this issue to gather use cases and +1s on a potential feature that would introduce instance store preservation for ECS Managed Instances. If this feature would be helpful for your workloads, please +1 and share your use case, instance types, and approximate cache sizes. This signal helps us prioritize.

**Attachments**
If you think you might have additional information that you'd like to include via an attachment, please do - we'll take a look. (Remember to remove any personally-identifiable information.)

No repository file, test, or implementation entry point is identified. Start by reviewing the ECS Managed Instances service and the instance-store lifecycle described in the issue. Done would require an agreed feature design and implementation scope for preserving local NVMe data.

Written by the indexing model from the issue text.