Update windows open queue to windows11 and VMimage to windows 2025

[LoopedBard3]

Update windows open queue to windows11 and VMimage to windows 2025.

Internal (successful) test run: https://dev.azure.com/dnceng/internal/_build/results?buildId=2731529&view=results

[LoopedBard3]

The internal run was successful.

[Copilot]

Pull Request Overview

This PR updates the build machine configuration to use the latest Windows images and queue names.

• Updated vmImage from windows-2019 to windows-2025
• Updated osVersion from 22H2 to 24H2 for public jobs
• Updated queue name from Windows.10.Amd64.Client.Open to Windows.11.Amd64.Client.Open

[LoopedBard3]

All of the scenarios failures are due to Process xxxx failed to run. Result: CloseFailed, the sdk_scenario failures are mostly due to Process 5220 failed to run. Result: TimeoutExceeded with some due to PermissionError: [WinError 32] The process cannot access the file because it is being used by another process: 'app'. The internal run on actual machines was successful so I am going to try a public run with windows-2022 instead of 2025 to see if that helps.

[LoopedBard3]

This test run that uses windows-2022 instead of 2025 still hits the same issue. https://dev.azure.com/dnceng-public/public/_build/results?buildId=1070518&view=results. I think it would make sense to take the update so we don't get caught with the old image deprecated and then spend some time investigating these specific cases.

[DrewScoggins]

Do we see these same failures on runs that are not a part of this PR?

I just looked and it seems like we are seeing the same failures. I think merging here is fine.

[LoopedBard3]

It seems the performance-ci scheduled run hits the same sdk_scenarios scenarios while the scenarios succeed, but the scenarios issues are not hit. So, it probably makes more sense to keep this until we either have a chance to investigate the scenarios or if we see mass failure due to the images going away.