rlxp-queue-batch

You can track RL experiment batch jobs without launching training by inspecting a local queue manifest to report ready, running, completed, and blocked jobs. This requires manifest state, contract approval, isolated job workspaces, and dry-run gating before any launch path.

A dry-run gated queue manifest is a local state file that tracks RLXP batch jobs across candidates, seeds, and phases. It keeps multi-run batches resumable and visible so jobs do not drift, ensuring no training launches until dry-run gating and contract approval are verified.

To identify which RL jobs are ready to launch next, inspect the queue manifest to mark jobs complete when expected outputs already exist. The manifest then summarizes blocked jobs and highlights what can launch next across smoke, short-trial, promoted, and incumbent-validation queues.

Yes, you can use queue batch tracking for smoke tests, short trials, promoted runs, and incumbent validation in task-scoped RL repositories. It operates across isolated job workspaces to keep multi-run RLXP experiment batches visible and resumable.

RLXP batch jobs show as blocked in the queue manifest when prerequisites like contract approval, isolated job workspaces, or expected outputs from prior phases are not met. The tracking system prevents launch paths until dry-run gating validates the required manifest state.