dos-supervise-loop

To keep dispatch workers alive, you need a fleet supervision process that continuously monitors lane leases and automatically replaces stalled workers. This approach maintains your target worker population by identifying missed ticks and filling gaps without disrupting healthy, advancing processes.

Lease-aware worker recovery is a dispatch loop mechanism that evaluates workspace lane leases to distinguish stalled workers from advancing ones. It reaps only stalled leases while surfacing spinning workers for human review, preserving healthy runs during fleet supervision.

You can supervise a DOS workspace fleet by applying a tick execution process that reads live taxonomy and kernel verdicts to manage worker rosters. This fills, scavenges, and surfaces workers across concurrent and exclusive lanes based on lease status and workspace availability.

Yes, you can recover from missed ticks without killing spinning worker processes. A lease-aware supervisor identifies the difference between stalled and spinning workers, reaping stalled leases to refill gaps while flagging spinning workers for human review rather than terminating them.

Yes, you need live taxonomy reads to run a dispatch-loop fleet supervisor. The supervision process requires live taxonomy data and kernel verdicts to execute lease-aware ticks, accurately fill population gaps, and scavenge stalled workers across the workspace roster.

You should not use a fleet supervisor for dispatch workers if your environment lacks lease-aware tick execution or live taxonomy reads. Without these kernel verdicts and lease monitoring capabilities, the supervisor cannot safely distinguish stalled, spinning, and healthy workers.