[BUG] Flowise Cloud Timeout limit issue
Describe the bug Flowise cloud api requests seems to time out around 60 seconds. This causes issues with agent flows which typically take longer to execute. To Reproduce Steps to reproduce the behavior:
Create any flow that will take more then 60 seconds to execute and try to call it through the api though postman or zapier. When I monitor it via langsmith the flow completes, but the connection always times out so I dont get the results returned.
Setup
Flowise cloud
Additional context Just about any decent agent flow takes around 60 seconds or more to fully execute so this seems like it may be a big usability issue for the cloud platform. I love the cloud option btw.
is the timeout happening everytime? Im guessing this is because the nginx proxt that we have. If possible, can you let us know how to replicate it?
Having the same issues, my agentflow is quite complicated and usually takes over 60 seconds to process After 60 seconds I always get following error:
504 Gateway Time-out
I have the same problem. And when I go to alter the timeout limit i see that is does nothing
Also have the same issue.
Particularly when using a tool calling agent and the tools each take time to generate a response to the agent.
I'm not entirely sure what the max timeout on a tool call is either. I've seen cases where the agent calls the tool again even though the first call may still be running - not ideal!
We just launched Flowise Cloud V2 at https://cloud.flowiseai.com, the issue should be resolved
I'm experiencing this right now using API to hit https://cloud.flowiseai.com/api/v1/prediction/. I see @vinicius-damin posted a screenshot of a timeout panel. How do you get to that?
@rhedayatpour click on "Additional Parameters" in the language model node
This continues to be a problem. I'm not sure if the setting actually does anything when using the API. It seems that API times out at 1 min no matter what. Also the vertex LLMs don't have this setting. How do I make sure that the API calls are not timing out at 1 minute? I had some success using streaming but in many cases my first streaming response can take more than a minute and the timeout triggers.
I am encountering the same 504 Gateway timeout when trying to upsert a pdf into pinecone vector db through upsert API endpoints. It was working fine previously but from today morning I am facing this issue. I am using the flowise from cloud.flowiseai.com.