Activepieces Not Working? Common Errors & Fixes

A webhook that works in test but not in production usually means the test URL was used on a published flow, the frontend URL is wrong, or the sender needs a different response. Use the live URL on a published flow and check the sender.

A run stays Queued when the project is at its concurrency limit, a burst filled the line, workers are busy or Redis lost data. Cancel unneeded runs, spread schedules out or add workers.

Flows and single actions time out after 10 minutes of active execution. Split the work into subflows or, on self-hosted, raise AP_FLOW_TIMEOUT_SECONDS.

Large lists or files can exhaust memory in a code step. Return only the data later steps need and process in batches.

A run is marked Succeeded when a step has Continue on Fail enabled, because the error becomes the step output. Inspect run details and branch on the output.

Most self-hosted failures come from a changed encryption key, no persistent volumes, a wrong frontend URL, an undersized server or exposed ports. Fix .env and server size first.

After self-hosted upgrades some users see empty flow lists or failing flows. Read the breaking-changes page, back up first and roll back by restoring the dump.

OAuth connections that fail or reset usually need deleting and re-creating in a clean browser session.

Flows that will not publish or that quietly stop usually have a broken trigger, connection or failed webhook registration. Re-check the trigger, reconnect and republish.

Cancel or downgrade from the Billing page in Activepieces Cloud, or email support if there is no self-serve option. Third-party guides, not the vendor, document these steps, so confirm them in your account.