Admin status

Included from Basic tier

The payload behind opschain admin status: GET /admin/status, also mounted at its admin alias GET /api/admin/status (both admin-token gated). One-look server health: build identity, which replica answered, dispatch quiesce state, scheduler membership and a queue summary. The same object is embedded verbatim as the server field of GET /api/admin/overview (shared builder).

The unprefixed /admin/* support paths (/admin/status, /admin/queue, /admin/licence, /admin/changes/active) share their exact paths with the embedded GUI's client-side admin routes. An unauthorised browser navigation to one — a GET carrying no Authorization header whose Accept prefers text/html, i.e. a refresh or deep link into the GUI — serves the SPA instead of the endpoint's 401, so GUI admin screens survive a refresh. Tooling is unaffected: any request presenting a bearer, or not asking for HTML, keeps the JSON endpoint's behaviour (401 on a bad or missing token), and gui-disabled or unbundled compositions keep the bare 401 everywhere.

The v2 OpenAPI document declares this resource's schema; the resource-field coverage gate holds this page to it.

started_at, draining, replica_id, leader and held_shards are per-replica facts: they describe the replica answering the request. The quiesce fields read the durable fleet posture and answer the same on every replica.

Fields

Field	Type	Meaning
status	string	ok, or degraded when the store ping (a limit-1 run list) failed — store then carries the error.
version	string	Build version; dev when unset.
commit	string	Build commit; dev when unset.
built_at	string	Build date; dev when unset.
started_at	timestamp	When this replica's process started.
replica_id	string	This replica's scheduler identity.
server_name	string	This node's stable server name (the server_name config key; default hostname) — the key of the per-node quiesce verbs.
site	string	This node's site identity (the site config key, e.g. dc1). Present only when configured (multi-site estates).
node_held	bool	This node is parked under an active server-node or site quiesce posture (held-at-boot included): readiness off, offers stopped, contending for nothing until the explicit resume.
leader	bool	Whether this replica is the scheduler's cron leader.
held_shards	int	Dispatch shards this replica currently owns.
draining	bool	This replica has begun shutdown.
quiesced	bool	Dispatch is held: by the durable fleet quiesce/drain posture (`POST /api/fleet/quiesce
quiesced_by	string	Who set the fleet posture, read from the posture row — the same answer on every replica. Present only while a fleet posture is active.
quiesced_at	timestamp	When. Same presence rule.
store	string	ok, or the store ping's error text.
identity	object	Present when the composition runs the credential stack (the PostgreSQL store). Its ldap_providers map reports each ldap-kind identity provider connection, keyed by connection key, as this replica has observed it through the credentials leg, the test verb and the sweep: configured, last_success_at (the most recent successful directory bind), and last_error/last_error_at (the most recent provider-unavailable fault, e.g. ldap directory unreachable at <host:port>: …). Absent fields mean no such outcome on this replica yet.
queue	object	Fleet-wide dispatch summary from the shared database: ready (parked entries) and leases (live job leases).
scheduler_replicas	array	Every scheduler replica in the shared membership table — fleet-wide, unlike the per-replica facts — each {"id": <string>, "heartbeat_age_seconds": <float>, "live": <bool>} plus "site" when the replica carries a site identity. Absent when the membership read returned none.

Where the values come from and what changes them

Build identity is baked at compile time; replica_id, leader and held_shards come from this replica's scheduler; draining flips at shutdown; quiesced/quiesced_by/quiesced_at from the durable fleet quiesce/drain/resume posture (plus this replica's shutdown hold); queue and scheduler_replicas from the scheduler's stats read over the shared database (a stats failure answers 500; a store ping failure only degrades status); identity.ldap_providers from the credential chain's own bind outcomes (per replica, in memory — a restart clears it; see opschain docs identity).

Related pages

• opschain docs fleet-queue — the per-entry view behind the queue summary, and the quiesce verb's effect on dispatch.
• opschain docs dispatch — shards, leadership and heartbeats.
• opschain docs identity — the provider connections behind identity.ldap_providers and the directory-outage behaviour.