D-179 worked example, stage 4 (the failure/stall handler). Lists codex_queue rows in state fail or fail_timeout, emails the failure kind (non-zero exit vs no evidence before the timeout - the silent/stalled case), and finalizes each row to failed_notified. File-free: the sweeper owns marker consumption. A late marker after a timeout would be reconciled by a later sweep, so an occasional follow-up success notification is expected (at-least-once by design). Run this as ONE scheduled dish - the scheduler serializes a schedule's runs, so the list-notify-mark sequence is not raced by a concurrent copy of itself.

How it works 8 steps

Inspect the data fetches, transforms, gates, and output this recipe runs.

Process (8 steps)

■ listing ?

■ rows filter

Filter by condition

⚠ guard_rows guard

Stop if rows is empty

■ now date_add

Add 0 days to

■ notify ?

■ mark_handled ?

■ handled_count count

Count items in rows

■ card to_summary

Format results as a summary card

Settings 2 configurable

Configurable at install. Defaults shown — change them anytime in Recued.

notify to setting = —

sender mail instance setting =