Skip to main content
Version: v5.6.1

opschain admin platform halt

Included from Basic tier

Emergency hard stop — the v2 PlatformService.HaltPlatform call. A durable per-plane posture row flips and every replica honours it:

  • Runner plane (--plane runners): dispatch freezes, queued work parks, leases are held — nothing ages into runner-lost — and in-flight steps are suspended in place (SIGSTOP / container pause) under --in-flight suspend, or run to completion under --in-flight complete.
  • Agent plane (--plane agents): channels severed, every agent credential refused.
  • Both (--plane both): full lockdown.

The halt is idempotent while active, and state is deliberately preserved for forensics: suspended attempts are disposed one by one with admin platform attempt (resume in place, or kill), and the plane comes back only through the separate, posture-row-named admin platform resume. The reason is recorded on the audit event.

Flags​

FlagMeaning
--plane <plane>Target plane: runners, agents or both (required).
--reason <text>Break-glass reason, recorded on the audit event (required).
--in-flight <mode>Runner-plane in-flight mode: suspend (default) or complete.
--server <url>Server base URL (default: the active context / OPSCHAIN_SERVER).
--auth-token <token>Admin bearer token override (default: your opschain login session, then $OPSCHAIN_ADMIN_TOKEN, then the token stored by opschain admin login).

Output is the server's response document (the new posture row, including the id a later resume must name) — JSON, or YAML under -o yaml.

Exit codes​

CodeMeaning
0Halt accepted.
1Request failed, the server refused the body, or the token was rejected.
2Usage error — missing required flags, or no server/credential resolvable.

Named errors​

  • a server is required — pass --server <url>, set OPSCHAIN_SERVER, or 'opschain context use <name>' — exit 2.
  • this command needs a login or an admin token — run 'opschain login', pass --auth-token, set OPSCHAIN_ADMIN_TOKEN, or store one with 'opschain admin login' — exit 2.
  • the server rejected the credential — run 'opschain login' again, or check the admin token — HTTP 401 (exit 1).

Examples​

opschain admin platform halt --plane runners --reason "suspected credential compromise"
opschain admin platform halt --plane both --reason "incident 4821" --in-flight complete

See also​

  • opschain docs cli/admin-platform-status — the active posture rows and history.
  • opschain docs cli/admin-platform-attempt — dispose suspended attempts.
  • opschain docs cli/admin-platform-resume — bring the plane back.
  • opschain docs cli/admin-server-quiesce — the non-emergency alternative when stopping new offers is enough.