Major issues with API today, timing out all of a sudden

I am wondering if there any current known issues? All of a sudden my web app stopped being able to call the API, nothing was changed in the code so I’m certain it wasn’t a change on my end that caused this. It was working fine then all of a sudden even with a 45 second timeout it still didn’t respond. I tried generating a new key, I was put through the wringer with arkose challenges to do this, I’m out of my home country currently, so not sure if that is why but that alone is very frustrating. With the new API key I am seeing the same issue.

I am not a heavy user of the API, my app is still in development so it’s only a handful of queries for testing. Very confused as to what is going on.

All good: https://llm-api-speed-meter.aicha.mp/

It is still quite slow for me, maybe my perception is just off as when I use the web app you can see it writing the response but I don’t have that turned on in my app, I will look at doing that and see if it seems comparable to chatgpt web app.

1 Like

Looks like rate limit cause your app not responding “suddenly”.

If you are in Tier1, check if you hit day token rate limit(check api response or account email notifications) when not respond as expected. When using assistant api with retrieval, it might cost more token than you expected in each api request, tense test/debug may hit rate limit easily, and it restore very very slowly. Upgrade your level may help, but it cost you more money.

If you want word by word stream response behavior, you cannot use assistant api for now and need to use chat completion api with stream=true, and handle server sent event. In my test, it did not cost extra tokens and harm rate limit(it produce more network traffic). Upgrade your level won’t give your api streaming behavior, it’s another story.

1 Like