[cloud.dify.ai] Node HTTP Request run failed: Reached maximum retries (3) for URL https://api.dify.ai/v1/chat-messages

[posmyk]

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

I use Node HTTP Request to connect to other Dify Application.
There is a POST to: https://api.dify.ai/v1/chat-messages but after first successfull request in first iteration the rest are timed out with the following error: Node HTTP Request run failed: Reached maximum retries (3) for URL https://api.dify.ai/v1/chat-messages",

2. Additional context or comments

Iteration:

http req:

Test run: Iteration 1 is OK, but the second fails with the timeout:

[MatheusBomtempo]

Did you find any solution to this problem?

[zhengjianping1010]

Did you find any solution to this problem?

[ccxuy]

Did you find any solution to this problem?
I tried #6244 , disabled SSRF_PROXY_HTTP_URL, SSRF_PROXY_HTTPS_URL, but still no luck.

[FrostbiteLee]

i meet the same problem.
i want to get the Knowledge Base file list from Node HTTP Request, url is "http://localhost/v1/datasets?page=1&limit=20", but failed.
When i change url to "https://httpbin.org/get", it's ok.

[Tenold-a]

I deployed Dify on a Windows 10 system using Docker Desktop. Initially, I attempted to send HTTP requests via a code node, but it kept failing and say: The requested URL could not be retrieved.

After checking the documentation, I discovered that Dify's sandbox environment restricts HTTP requests from the code.

Switching to the HTTP Node

Since the code node couldn’t send requests, I switched to Dify's HTTP node but encountered the same issue as you. I spent a lot of time troubleshooting, trying various fixes, including removing the SSRF_PROXY settings mentioned in (#6244), but nothing worked.

The Key

After extensive searching, I know that:

• Dify's code execution happens inside a sandbox container.
• The API service is hosted in a separate container.
• Containers in Docker can communicate with each other using container name + port number, as long as they are on the same internal network.

The Fix

I checked the Nginx configuration and in default.conf.template found that requests to localhost/v1 were actually being forwarded to http://api:5001.

    location /v1 {
      proxy_pass http://api:5001;
      include proxy.conf;
    }

So, I changed my HTTP request to:

http://api:5001/v1

It is work for me

[ocyisheng]

爱老虎油