Troubleshooting

Common errors with their actual cause and the fix. If you don't see yours here, open a ticket from the customer dashboard with the execution id and the error message from the timeline — that's the fastest path to a real answer.

"The workflow never fires"

Most common cause: the workflow is inactive. Check the status toggle at the top right of the editor — it should be green.

Second most common cause: the trigger references something that doesn't exist (e.g. a WooCommerce trigger on a site without WooCommerce installed). Open the workflow in the editor — the trigger node shows a red exclamation and an error in the inspector.

Third: WP-Cron is disabled. Schedule triggers and delayed executions need cron. Check with wp cron event list (wp-cli) or look at Setyenv → Settings → Cron health. Fix: configure a real external cron hitting wp-cron.php every minute, and define DISABLE_WP_CRON in wp-config.php.

"Executions are stuck in running"

The execution worker process hasn't picked it up yet, OR the queue is backed up, OR a long-running node is genuinely still running.

1. Check Setyenv → Settings → Queue health. Shows pending count, oldest pending age, last worker tick.
2. If oldest pending age is over 1 minute, the queue worker is likely not running. The queue is drained by either WP-Cron (default) or by a long-running PHP CLI worker (recommended for production). Bring up the worker.
3. Long-running nodes: AI nodes can take 30s+ for big completions, HTTP nodes can take 60s. The default timeout is 300s per node. Configure per-node in the inspector.

"A node fails with connection_not_found"

The node references a connection that was deleted or whose id changed (after an import). Edit the node in the editor and re-pick the connection from the dropdown.

"An HTTP node returns 401 / 403"

The credential the node is using has been rotated or revoked at the third party. Edit the connection and re-fill the credential. Use the Test connection button before saving to confirm.

"An AI node returns rate_limit_exceeded"

Your provider's API is throttling you. Wrap the AI node in a Retry block with exponential backoff. The default 3 attempts × 2.0 factor usually absorbs transient bursts.

For sustained traffic over the rate limit, upgrade your provider plan or distribute traffic across multiple keys (use a Condition node + multiple AI nodes with different Connections).

"A WooCommerce trigger fires twice for the same order"

Almost certainly a third-party plugin re-firing the WooCommerce hook. The engine's idempotency already protects against double execution — the second event is detected and the original execution's result is returned. If you see TWO execution rows for the same order in the executions list, one of them will have status idempotency_hit and zero duration.

If you want to opt out of idempotency on a specific trigger (very rare), set idempotency: false on the trigger node config.

"The editor canvas is laggy with my big workflow"

Browser rendering of 500+ nodes hits a frame-rate wall. Split the workflow into sub-workflows called via Call function — both faster to render and easier for humans to read.

"I get capability_denied when calling the REST API"

The Application Password (or the cookie-auth user) doesn't have the required capability. Map the capability to the user's role via the WordPress User Role Editor plugin, or assign the user a role that already has it (admin has all by default).