JobMonitor

Background job queue dashboard. Polls a jobs resource (or whichever you pass), groups rows by status, and exposes retry / cancel actions per row.

Installation

pnpm dlx shadcn@latest add @shadmin/job-monitor

Usage

import { JobMonitor } from '@/components/admin';
 
<JobMonitor resource="jobs" pollInterval={5000} />
<JobMonitor
  resource="jobs"
  pollInterval={2000}
  tabs={["failed", "running", "queued"]}
/>

Job record shape

interface JobRecord {
  id: string | number;
  type?: string;
  status: "queued" | "running" | "failed" | "done" | "cancelled";
  payload?: unknown;
  attempts?: number;
  lastError?: string | null;
  createdAt?: string;
  updatedAt?: string;
}

Props

Prop	Required	Type	Default	Description
`resource`	Optional	`string`	Context	Override resource
`pollInterval`	Optional	`number`	`5000`	Polling interval in ms (`0` = disabled)
`tabs`	Optional	`JobStatus[]`	`["running","queued","failed","done"]`	Tab order
`statusSource`	Optional	`string`	`"status"`	Record field name
`typeSource`	Optional	`string`	`"type"`	Record field name
`payloadSource`	Optional	`string`	`"payload"`	Record field name
`errorSource`	Optional	`string`	`"lastError"`	Record field name
`attemptsSource`	Optional	`string`	`"attempts"`	Record field name
`actions`	Optional	`boolean`	`true`	Show retry/cancel buttons

Actions

Retry — visible on failed rows. Sets status: 'queued' and increments attempts.
Cancel — visible on running and queued rows. Sets status: 'cancelled'.

Polling

Wraps useGetList with TanStack Query's refetchInterval. Set pollInterval={0} to disable. Recommended values: 2000–10000 ms.

Limitations (v1)

No log-tail (would require a separate streaming endpoint).
No payload editing — payload is read-only.
No bulk-retry / bulk-cancel across many rows.
Cache hit/miss tracking deferred to <DataProviderDevtools>.