AEM workflows fail with “Process implementation not found” after author pod restart
This article explains how to troubleshoot workflow failures in AEM as a Cloud Service when the error Process implementation not found appears after an Author pod is recycled or restarted. The issue affects custom and out-of-the-box workflow steps and can also make the Developer Console inaccessible. To fix this, review the error logs, have the affected Author pod restarted or replaced through the Adobe-managed recovery process, wait for the workflow services and bundles to initialize, and resume the affected workflows.
Description description
Environment
- AEM as a Cloud Service (all supported versions)
- Author environment using containerized or cloud-native infrastructure, such as
Kubernetespods
Issue/Symptoms
- Workflow instances fail with the error com.adobe.granite.workflow.WorkflowException: Process implementation not found:
[process class name]. - The error appears in the AEM logs after an Author pod is recycled or restarted.
- Both custom and out-of-the-box workflow steps are affected.
- The issue is reproducible when you run an affected workflow on an asset.
- The Developer Console becomes inaccessible and redirects you to the login page after successful authentication.
Cause
When workflow jobs are picked up during an Author pod recycling or restart event, workflow process implementations becomes temporarily unavailable. Restarting or replacing the affected Author pod reinitializes the required workflow services and bundles and restores workflow processing.
Resolution resolution
Follow the steps below to resolve the issue:
- Check whether the workflow failures coincide with a recent Author pod recycling or restart event.
- Review the AEM error logs for entries similar to com.adobe.granite.workflow.WorkflowException: Process implementation not found:
[process class name]. - If the issue occurs during or after an Author pod recycling event, contact Adobe Support to request a restart or replacement of the affected Author pod so that the workflow-related services and bundles can initialize properly.
- Wait for the Author pod to become fully available and confirm that the workflow services are registered and active.
- Rerun or resume workflow instances that remained incomplete during the incident window.
- Monitor workflow execution to confirm that the error doesn’t recur.
- If the issue recurs, record the approximate timestamp and the affected workflow or asset, and provide those details to Adobe Support for further investigation.
Note: Author pod restart or replacement is a platform-level recovery action. Use Adobe Support rather than attempting to manage the underlying pods directly.
Related reading
Workflows remain stuck in the Running state due to an author pod deadlock