Day-two commands

The commands you reach for once it is running.

hookspool status                     # the one-screen health view (same JSON as /healthz)
hookspool issues                     # open issues; ack / resolve / ignore <id>
hookspool list --state dead          # then: show <id> --events, split, exclude, drop
hookspool replay --dead --destination hermes:github-ci   # retry dead batches after the cause is fixed
hookspool pause hermes:github-ci --reason "gateway upgrade"   # destinations only; resume undoes it
hookspool export --since … > backup.jsonl                     # forensic copy of batches, events, attempts

Back up the state directory with SQLite's own copy (sqlite3 spool.db "VACUUM INTO '…'") or hookspool export, not by copying spool.db alone while the dispatcher runs. Details are in Architecture, sections 8, 12 and 16. Take one before an upgrade that carries a migration.