Jobs

Almost everything slow on the platform runs as a background job: provisioning a server, a wipe, a plugin install, a ban sweep. All jobs (Infrastructure group) is the platform-wide queue with the schedules that create jobs. Capability: operations. Community admins see only their own community's jobs in their panel.

Queue health

A strip at the top shows the state of the job system:

Item	Meaning
Leader	The worker that runs schedules ("none" in amber means schedules are not running).
Workers	Number of active workers (amber at 0).
Counters	Jobs per status.
Oldest queued	Shown when a job has waited more than 5 seconds.

Auto-refresh reloads every 5 seconds; Refresh reloads now.

Jobs tab

Filter by status with the chips and by type with the text box (press Enter).

Status	Meaning
queued	Waiting for a worker.
running	In progress.
waiting	Waiting on a server agent (an "agent unreachable" note shows when it cannot be reached).
succeeded	Done.
failed	An attempt failed; it may be retried automatically.
dead	All attempts used.
cancelled	Cancelled.

The table lists status, type, target, attempts (used out of maximum), run time and finish time, 25 per page. Retry appears on failed, dead and cancelled jobs; Cancel on queued, running and waiting jobs.

Click a row for the detail drawer: ID, attempts, whether the job is idempotent, target, concurrency and dedupe keys, times, who holds it, progress, the error (code, message, stack), the last 50 log lines, the payload and the result. Agent jobs show live steps from the server's agent.

Schedules tab

All recurring work: platform tasks (ban sweep, VM health, upstream checks), wipes, restarts and Discord posts of every server.

Control	Detail
Group chips	All, Wipes, Discord, Platform and other groups.
Server	A server, or Platform-wide only.
State	Any state, Failing, Overdue, Disabled.
Search	Name or description.

Columns: name (with a disabled tag and a description tooltip), server, schedule in words and the cron expression (UTC), Next run (with overdue when it is late, which usually means the scheduler is not running), Last run (status, time, error), Failures (consecutive) and History.

History opens the run history: every failed run (identical consecutive errors are grouped with a count) plus the last 3 successes, each with View job.

Times are shown in your local time zone; cron expressions are UTC.

Troubleshooting

Problem	Fix
Jobs stay queued	No worker: check Workers and Leader in the health strip.
A job is dead	Read the error and logs, fix the cause (VM, server, plugin), then Retry.
Schedule overdue	The leader is missing, or a long job blocks it; check the health strip.
A wipe did not start	Look at the schedule's History and the matching job.