PushPals CLI Release Log

Release Metadata

Fail orphaned claimed jobs as soon as the server can prove a worker heartbeat has dropped ownership, instead of leaving the job stuck in claimed until the stale-claim watchdog fires minutes later.
Add an explicit abandoned recovery state for retry-safe lost claims and preserve successor lineage through resumeOfJobId, attempt tracking, and params.resume metadata instead of flattening every recovery into a hard failure.
Auto-requeue only retry-safe work such as warmup.execute, while keeping non-idempotent task.execute jobs as hard failures with clearer diagnostics and updated queue-health telemetry.
Update server and dashboard surfaces to expose abandoned jobs in queue snapshots, lifecycle timing, and failure-rate style health signals, with regression coverage for both heartbeat-mismatch and stale-watchdog recovery paths.

npm i -g @pushpalsdev/cli

bun install -g @pushpalsdev/cli

task.execute resume metadata is preserved for future continuation logic, but non-idempotent abandoned jobs still restart via explicit retry rather than automatic in-place continuation.

Full Changelog: v1.0.56...v1.0.57