Request limit/TPS
I recently encountered an issue while processing a batch test of 170 people through a Typebot automation and it froze several times. To better understand and optimize the system, I wonder if anyone can provide information on the system's capacity limitations, specifically in terms of Transactions Per Second (TPS) or any request limits that might be in place? I want to ensure that the system can handle the desired load without encountering performance issues.
6 Replies
What do you mean by « froze »? What do you see exactly? What do you monitor?
This is the video of a user trying to enter his name.
Under the name block I had a OpenAi request to check the user gender using the user's name. It would not go through.
In the end of the automation the bot also wouldn't load and the same thing would happen, the user would press "continue" and the automation would stall and go back to the "continue" button
Have you checked the logs for that particular session? https://docs.typebot.io/results/overview#logs
Typebot Docs
Results - Typebot Docs
I did, I got a 429 error from OpenAI request, which is the too many request per minute. However, I need to brief the CEO this week and I want to, in addition to showing the error I got from OpenAi, show what are the TPS on typebot servers. Can you provide that information?
Hmmm I don't have that info for Typebot Cloud. But it gets a lot of requests 😅à
So you've got the issue, it's not a load problem
Just a hitting rate limit on OpenAI's side
Thank you Baptiste!