Help：While I call the LLM node, it always return this error in docker-api-1 logs. What's the reason?

[simonjhy]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:)
• Please do not modify this template :) and fill in all the required fields.

1. Is this request related to a challenge you're experiencing? Tell me about your story.

While I call the LLM node, it always return this error in docker-api-1 logs. What's the reason? Its the LLM reply time too long? How to configure the node "Read timed out" time_out value?

2025-12-19 11:58:19.656 ERROR [GraphWorker-1] [base.py:277] - Error in stream response for plugin {'code': -500, 'message': '{"message":"{\"args\":{},\"error_type\":\"ConnectionError\",\"message\":\"HTTPConnectionPool(host='192.168.67.216', port=8888): Read timed out.\"}","error_type":"PluginInvokeError","args":null}', 'data': None}
2025-12-19 11:58:19.658 ERROR [GraphDispatcher] [error_handler.py:101] - Node 1766031675572 failed with ABORT strategy: req_id: e3ca58b7af PluginInvokeError: {"args":{},"error_type":"ConnectionError","message":"HTTPConnectionPool(host='192.168.67.216', port=8888): Read timed out."}
2025-12-19 11:58:19.672 INFO [GraphWorker-1] [_client.py:1038] - HTTP Request: POST http://plugin_daemon:5002/plugin/7b6b653b-28e2-463a-bdd2-219ad621abd1/dispatch/llm/invoke "HTTP/1.1 200 OK"
2025-12-19 11:58:19.676 ERROR [GraphWorker-1] [base.py:277] - Error in stream response for plugin {'code': -500, 'message': '{"message":"{\"args\":{},\"error_type\":\"ConnectionError\",\"message\":\"HTTPConnectionPool(host='192.168.67.216', port=8888): Read timed out.\"}","error_type":"PluginInvokeError","args":null}', 'data': None}
2025-12-19 11:58:19.729 ERROR [GraphDispatcher] [error_handler.py:101] - Node 1766031675572 failed with ABORT strategy: req_id: 225dc8f3b8 PluginInvokeError: {"args":{},"error_type":"ConnectionError","message":"HTTPConnectionPool(host='192.168.67.216', port=8888): Read timed out."}
2025-12-19 12:02:29.030 ERROR [GraphDispatcher] [error_handler.py:101] - Node 1766031634703 failed with ABORT strategy: req_id: 7cc6778f79 PluginInvokeError: {"args":{},"error_type":"ConnectionError","message":"HTTPConnectionPool(host='192.168.67.216', port=8888): Read timed out."}
2025-12-19 12:02:29.130 ERROR [Thread-910 (_generate_worker)] [workflow_entry.py:111] - Unknown Error when workflow entry running
Traceback (most recent call last):
File "/app/api/core/workflow/workflow_entry.py", line 107, in run
yield from generator
File "/app/api/core/workflow/graph_engine/graph_engine.py", line 270, in run
raise self._graph_execution.error
RuntimeError: req_id: 7cc6778f79 PluginInvokeError: {"args":{},"error_type":"ConnectionError","message":"HTTPConnectionPool(host='192.168.67.216', port=8888): Read timed out."}

2. Additional context or comments

No response

[dosubot[bot]]

The "Read timed out" error happens because the LLM node or plugin daemon isn't responding within the configured timeout window. This can be due to the LLM reply taking too long, network issues, or resource exhaustion on the LLM node or plugin daemon.

You can increase the timeout by setting the PLUGIN_DAEMON_TIMEOUT environment variable (in seconds) in your .env file or docker-compose.yaml. For example, to set a 20-minute timeout, add:

PLUGIN_DAEMON_TIMEOUT=1200

After updating the value, restart your Docker containers for the change to take effect. The default is now 600 seconds (10 minutes) in recent Dify versions, but you can set it higher if your LLM responses are slow or you have long-running tasks. This setting directly controls how long Dify waits for a response from the LLM node before giving up and throwing a timeout error. More details and code references are available in the docs and recent PRs, and you can see how the timeout is used in the plugin client code.

If increasing the timeout doesn't help, check your network connectivity, make sure the LLM node is healthy and not overloaded, and review logs for any other errors. Other related settings you might want to look at include PLUGIN_MAX_EXECUTION_TIMEOUT and HTTP_REQUEST_MAX_READ_TIMEOUT, depending on your workflow and where the timeout is occurring.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[simonjhy]

I did noit find the PLUGIN_DAEMON_TIMEOUT in .env file. Which docker service I need to configure? docker-api or web in docker-compose.yaml?

[dosubot[bot]]

You should set PLUGIN_DAEMON_TIMEOUT for the docker-api service, not the web service. If it's missing from your .env file, you can add it manually (for example, PLUGIN_DAEMON_TIMEOUT=1200 for 20 minutes), or set it directly in the api service's environment section in your docker-compose.yaml. After making changes, restart the docker-api container for the new timeout to take effect. The web service does not use this variable—only the API service does for plugin/LLM node communication timeout control.
Reference: docker-compose.yaml

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}