The workflow gets stuck at a certain node, which remains in a running state. When there are many nodes, the nodes after concurrency always show a running status, causing the process to stall.
1 Like
Encountered the same issue, the workflows_run table remains in a running state, seemingly causing a queue blockage.
I’ve also encountered this issue. You can check the server logs:
Scenario 1: The selected LLM model is rate-limited by the provider (e.g., TPM, RPM).
If this is the case, log in to the provider’s platform (such as Alibaba Cloud Bailian) to view the model’s rate-limiting statistics and contact them to request scaling up.
Scenario 2: Caused by Dify execution timeout.
If this is the case, adjust the timeout limits accordingly, for example:
APP_MAX_EXECUTION_TIME=3600 (original value: 1200)
WORKFLOW_MAX_EXECUTION_TIME=3600 (original value: 1200)
I updated to 1.11.3 and solved the node running issue.