Monitoring Jobs

Everything about a job after you create it lives on the same AI Product Tools > AI Automation Jobs page, including the one you just saved. This page covers the list, a job's own detail, its logs, and the actions that pause, resume, retry, or remove it.

The job list

Every job you have, across every status, appears in one table: name, tool, status, progress, created date, and next run.

Statuses

WP-Cron health banners

Automation runs entirely on WP-Cron, so the list page watches for two different signs of trouble:

• Confirmed: shown only when DISABLE_WP_CRON is set to true in wp-config.php. The banner names the fix directly: set it back to false, or configure a real server cron hitting wp-cron.php.
• Heuristic: shown instead, only when the confirmed case does not apply, if a job that should already be running shows no sign of having started more than 2 minutes past its due time. This is not proof of a server problem, since WP-Cron only fires on incoming site traffic and a quiet site can delay it on its own; the banner offers "Run jobs in this browser now", which polls from your open tab as a stopgap while you leave it open.

Neither banner changing color or disappearing is itself proof a job is now running; check the job's own row. See Background jobs for the full WP-Cron diagnosis, including your host's cron setup.

A third, separate surface can also appear on this page: if WP-Cron is confirmed disabled (the same condition as the banner above) while any job is still active, pending, scheduled, preparing, or running, a "Background Processing Required" dialog pops up on its own, without you clicking anything.

A job's own detail

Click anywhere on a job's row (its name, status, or dates, just not the checkbox or the actions on the right) to expand its detail inline, right below the row.

The stat row repeats progress, product count, success, and failure counts; Edit, Duplicate (creates a copy named after the original with " (Copy)" added), and Delete sit to the right. A Retry Failed (n) button, where n is the failed count, appears once the job has at least one failed product; it creates a brand new job targeting only the products that failed, rather than re-running the original.

Per-product results

The Products panel lists every product the job touched or will touch, with filter chips for All, Pending, Processing, Completed, and Failed.

A product that keeps failing with the identical message across every attempt points to a real, ongoing provider problem (a bad key, no quota, a retired model), not a fluke; see Provider errors for what each exact message means.

Logs

Open View Logs (from the row's More actions menu, shown below) for a complete, timestamped run log.

Logs are kept for 30 days by default and then pruned automatically; each entry is capped at 4 KB of message text, so copy anything you need before it ages out.

Controlling a job

Stopping a job is immediate and does not ask you to confirm which products to keep: whatever had already been written stays written, and whatever had not started is simply never attempted.

A new job does not start the instant you create it, and neither does a manual Run Now or Trigger Now: a brand new or recurring-cycle job schedules its first tick about 10 seconds out, while a manual Run Now or Trigger Now schedules its tick about 5 seconds out. Either way, each tick processes up to 10 products in roughly 25 seconds before scheduling the next one 2 seconds later. At most 3 jobs run at the same time site-wide; any more wait their turn in Waiting to run.

Troubleshooting