The website now publishes the approved direction for independent CPU, RAM and temporary disk selection. Core, API and portal support will follow.

  • Customers will be able to save reusable resource profiles and copy their workflow label. Existing runner labels remain supported. The new independent selection is not available yet.
  • The approved scheduling model uses overcommit-aware allocatable capacity, host headroom and resource pressure, with safe memory and disk reservation and fair queueing. Arbitrary plan-based concurrency caps are not part of that target. Current service limits remain in effect until rollout.
  • Flexible profiles will draw from the existing unified equivalent-minute balance. New CPU, RAM and disk rates have not been announced. The pricing page keeps current standard-runner prices separate from the approved direction.
  • Read the Flexible runner profiles guide for approved resource ranges, capacity states and rollout requirements. This announcement publishes the design, not a completed service rollout.