Ability to select runner region in multi‑region high‑availability setup #4385
Description
What would you like added?
We would like a way to select or target the region of a self‑hosted runner when using a multi‑region high‑availability setup (as described in #4283).
Currently, runners across regions share the same scale set name for HA purposes, and we don’t have a method to apply additional labels that would differentiate runners by region. A mechanism to specify runner region in workflow configuration - or a supported pattern to assign region-specific labels - would allow us to direct certain workflows to a specific region when necessary.
Why is this needed?
Although most of our workflows perform equally well regardless of runner region, there are a few specific cases where the execution region has a measurable impact.
We explored using scale set names or custom labels, but the HA design requires identical scale set names across regions, and we do not currently have a supported way to apply region‑unique labels. As a result, there is no current method for our workflows to choose the region of the underlying runner.
Additional context
We have successfully implemented multi‑region runner high availability and it works extremely well. The only missing capability is region targeting for a subset of specialized workflows.
Any guidance or recommended patterns would be greatly appreciated.