.NET: support hierarchial partition key on CosmosDB Actor Storage

[Korolev Dmitry (DeagleGross)]

Support more complex partition key on cosmosdb. Updated packages as well.

Also fixes#356 where test projects for CosmosDB were removed. I've added them back (recreated to make build on new VS work) and also set the CosmosDB emulator in the CI.

Fixes#305

Contribution Checklist

• The code builds clean without any errors or warnings
• The PR follows the Contribution Guidelines
• All unit tests pass, and I have added new tests where possible
• I didn't break anyone 😄

[Copilot]

Pull Request Overview

This PR introduces support for hierarchical partition keys in CosmosDB Actor Storage by changing from a single partition key (/actorId) to a composite partition key using both /actorType and /actorKey. The changes also update Aspire packages to their latest versions.

• Implements hierarchical partition key support for better data distribution and query performance
• Updates composite indexing to align with the new partition key structure
• Updates Aspire package versions to 9.4.1

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File	Description
LazyCosmosContainer.cs	Updates partition key configuration from single /actorId to hierarchical /actorType and /actorKey, and adjusts composite indexing accordingly
CosmosTestFixture.cs	Updates test fixture to use the new hierarchical partition key configuration
Directory.Packages.props	Updates Aspire package versions from 9.3.1/9.4.0 to 9.4.1

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Aditya Mandaleeka (adityamandaleeka)]

Aspire 9.4 for the win! Looks good, is this something we can test at all or does the SDK abstract this away entirely when you're querying?

[Korolev Dmitry (DeagleGross)]

Isthis something we can test at all or does the SDK abstract this away entirely when you're querying?

I dont think there is an easy way to test hierarchial partition keys in terms of performance: we would need to generate synthetic data; fill it in db and do the perf queries - so I think we can avoid it at this moment honestly.

SDK is abstracting the partition keys, because it allows querying and posting data as before without any changes.

[Korolev Dmitry (DeagleGross)]

skipping 3 tests from the cosmos suite - the tests are stable and passed the run for 3 times in a row already.
Ready to be merged

[Korolev Dmitry (DeagleGross)]

is a part of #402 as well and merged already!