Thanks again for raising this and for confirming that auto-cancellation would work for you, so I'm closing the request for now.
Black Breeze
•
Sep 1, 2025
Thanks for this feedback!
You've identified a real pain point with module test capacity management, but my worry is that manual cancellation would create poor UX. The core issue isn't that users can't cancel tests - it's that they shouldn't need to.
When you're rapidly iterating on a module, older queued tests become obsolete and just waste worker capacity. Making users manually manage this queue turns them into "human process schedulers."
Can we please explore options that don’t put extra operational burden on you as a user? Wouldn’t auto-cancellation (default or even policy-based) work better for you? Are there any scenarios where you’d still require this level of manual control?
Fuchsia Backpack
•
Sep 5, 2025
Yes I think that would work. Only assumed I’d have to write a policy because for stacks this isn’t a built in feature.
Log in to comment and vote
Comments3
Aleksandra Cieslak
Oct 29
Thanks again for raising this and for confirming that auto-cancellation would work for you, so I'm closing the request for now.
Black Breeze
Sep 1, 2025
Thanks for this feedback!
You've identified a real pain point with module test capacity management, but my worry is that manual cancellation would create poor UX. The core issue isn't that users can't cancel tests - it's that they shouldn't need to.
When you're rapidly iterating on a module, older queued tests become obsolete and just waste worker capacity. Making users manually manage this queue turns them into "human process schedulers."
Can we please explore options that don’t put extra operational burden on you as a user? Wouldn’t auto-cancellation (default or even policy-based) work better for you? Are there any scenarios where you’d still require this level of manual control?
Fuchsia Backpack
Sep 5, 2025
Yes I think that would work. Only assumed I’d have to write a policy because for stacks this isn’t a built in feature.