The job of this console
See what is failing, trigger a safe manual run when operations need it, and leave timing to the schedule. Do not hammer Run now on payroll-adjacent jobs without knowing the window.
Admin · Operations
Scheduled Jobs
ScheduleType inventory with last-run status and manual Run now / Run scan.
Scroll to Flowcharts. Scenario: Run a scheduled job on demand.
Scheduled work
Last run, next run, and a careful Run now
Scheduled jobs are the ScheduleType inventory with last-run status and manual Run now / Run scan. Accruals, reminders, and scans live here; they are not started from a hidden cron UI.
Admin → Scheduled jobs (deep link)
Route
/admin/jobs
HR Admin · Org Admin
Seat at the table
Administrators and support engineers. Employees never run jobs. Developers still own job code; this is operations, not a code editor.
The working rhythm
Scan last-run status, open a failing type, read the error, fix source data or config, then Run now if the job is idempotent enough. Export only for incident notes.
Moments that bring you here
Leave accrual did not fire. A reminder storm. After a flag change that a job gates. Overnight failure in the admin dashboard job card.
ⓘ
Jobs mutate source records on purpose
A run writes to the domain it serves (balances, notifications, scans). Treat Run now as a production action, not a refresh button.
Flowcharts
1 — Run a job now
Find a ScheduleType, inspect last run, then trigger Run now.
flowchart TD
A[Open Scheduled jobs] --> B[Review last-run]
B --> C[Select a type]
C --> D[Run now]
D --> E[Watch status]
E --> F[Inspect errors if failed]
- 1Open jobsAdmin → Scheduled jobs (deep link).
- 2Inspect last runStatus and timestamp per ScheduleType.
- 3Run nowUse for recovery, not as a replacement for cron.
- 4Confirm effectCheck the domain screen the job feeds.
Overview
Jobs operate on existing domain tables (timesheet lock, leave accrual, retention purge). This page is an operations console, not a new scheduler database.
How to use
1. Do not hammer Run now
Overlapping runs can lock rows.
2. Read the type name
ScheduleType maps to a known backend job.
3. Failed runs
Fix data or flags before retrying.