Add azurelinux.3.cobalt.arm64.perf queue to runs

[LoopedBard3]

Add azurelinux.3.cobalt.arm64.perf queue to dotnet perf slow pipeline and private sdk testing runs.

dotnet-runtime-perf-slow test run: https://dev.azure.com/dnceng/internal/_build/results?buildId=2861739&view=results
dotnet-performance test run: https://dev.azure.com/dnceng/internal/_build/results?buildId=2861740&view=results

[Copilot]

Pull request overview

This PR adds support for the Azure Linux 3 Cobalt ARM64 performance queue across the pipeline configuration. The changes enable performance testing on this new machine type for both SDK and runtime scenarios.

• Adds perfcobalt logical machine mapping to the azurelinux.3.cobalt.arm64.perf queue
• Extends log redaction support to include Azure Linux OS group
• Adds new build matrix template for azurelinux3-arm64-cobalt machines
• Integrates Cobalt queue into SDK performance jobs and runtime slow performance jobs

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
scripts/run_performance_job.py	Maps the perfcobalt logical machine name to the Azure Linux 3 Cobalt ARM64 performance queue
eng/pipelines/templates/run-performance-job.yml	Extends log redaction task to support Azure Linux OS group alongside Ubuntu
eng/pipelines/templates/build-machine-matrix.yml	Adds build matrix configuration for Azure Linux 3 ARM64 Cobalt machines
eng/pipelines/sdk-perf-jobs.yml	Adds azurelinux3-arm64-cobalt to multiple SDK performance job configurations
eng/pipelines/runtime-slow-perf-jobs.yml	Adds performance jobs for Mono interpreter, JIT, AOT, and CoreCLR on the Cobalt queue

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[DrewScoggins]

Any idea why the runs were skipped in the perf-slow?

[DrewScoggins]

Also, since we only have five machines in the queue currently, we should only have one perf-slow run in total. At least until we increase the queue size.

[LoopedBard3]

Any idea why the runs were skipped in the perf-slow?

I still need to make changes for the install prereqs steps so I cancelled the currently linked run.

Also, since we only have five machines in the queue currently, we should only have one perf-slow run in total. At least until we increase the queue size.

Any preference on the run to do? Coreclr micro is what I would think.