4 months ago
Hello,
We have an instance that calls OpenAI very often. I very commonly observer incredibly slow response times. Same requests on local machines are about 10x faster, and I can't seem to observe any problems on the OpenAI side. I am very confused around how to start debugging this issue and would love some help.
3 Replies
4 months ago
Hey there! We've found the following might help you get unblocked faster:
If you find the answer from one of these, please let us know by solving the thread!
4 months ago
This thread has been marked as public for community involvement, as it does not contain any sensitive or personal information. Any further activity in this thread will be visible to everyone.
Status changed to Open brody • 4 months ago
4 months ago
I am increasingly more confident that I am doing something wrong with concurrency and networking. Would love some help in understanding the problems
aysan0
I am increasingly more confident that I am doing something wrong with concurrency and networking. Would love some help in understanding the problems
4 months ago
To help narrow this down, could you share a bit more about your setup?
What language/SDK are you using to call the OpenAI API?
What kind of container is your Railway instance running on? (If it’s a private image, a rough description is fine.)
Do you have a minimal code sample or a Railway project where the slowdown can be reproduced?
Which OpenAI API endpoint(s) are you calling (e.g. chat/completions, responses, embeddings)?
Roughly how slow is it on Railway compared to local? And how much faster are the same requests when run locally? Please include approximate timings in seconds if possible.
For reference: I’m using Node.js with the OpenAI Responses API, and haven’t noticed unusual slowness on Railway. I’ve tested in both Singapore and US West regions.
This info would really help in figuring out whether the slowdown is specific to your setup or something in the network path.