We have seen a lot of API timeout error while using gpt-3.5-turbo model as below in the last few days
ReadTimeout: HTTPSConnectionPool(host='api.openai.com', port=443): Read timed out. (read timeout=600)
Anyone else also got a lot of timeout error?
The code we used is like this
completion = openai.ChatCompletion.create(
model="gpt-3.5-turbo",
temperature = 0.1,
n = 3,
messages=[
{"role": "system", "content": system_content},
{"role": "user", "content": user_content}
]
)
Yes, lately it’s been hit-or-miss.
I completely recommend using a backoff library. I like retry · PyPI
You can even hook a “We’re sorry, it appears the server is currently overloaded” message after so much time
They also have a status page: https://status.openai.com/
Although, it does seem delayed.
Been receiving a bunch of errors and timeouts today as well. Did not experience this last week.
Also getting 502s randomly.
raise error.APIError(
openai.error.APIError: HTTP code 502 from API (<html>
<head><title>502 Bad Gateway</title></head>
<body>
<center><h1>502 Bad Gateway</h1></center>
<hr><center>nginx</center>
</body>
</html>
)
Yeah, everyone is seeing the same “timeouts” and across both chat completion and completion API methods and models.
Strangely, the OpenAI status page shows everything “green”:
HTH
Does OpenAI know about this? Who should we tag?
I am sure they know about it @shuo
They have over 100 millions users all over the globe.
It’s best to relax and wait for the issue to clear, to be honest.
HTH
I’ve also been getting ECONNRESET and Request failed with status code 502 errors in the past 10 hours. BTW openai api takes several minutes before throwing an error and timing out, which is a bit annoying.
You can write code in your app to shorten this timeout.
Normally, this is a good time to work on error and exception handling code when performance is bad.
For example, check out this current topic:
Makes sense. I just set a 30 seconds timeout and then I try with gpt-3.5-turbo-0301. Not sure if changing the model will make any improvement though
having 3 + days of failed requests because of timeouts is pretty bad for a business
same error to me, what is the solution? 
It does. Sometimes, outages do not work the same way for all the models/endpoints (sometimes they do).
@ljia31523 All the strategies are pretty much explained in this link that @ruby_coder already shared. Feel free to reach out if you still have doubts or do not know how to implement them.
Same problem with davinci-003 and 3.5-turbo
Experiencing the same issue with both davinci-003 and davinci-002.
It’s almost been 3 days with the same issue. It’s making it really hard to work.
what is most annoying to me is:
I try to set a timeout at openai.ChatCompletion.create(..., timeout=10) hoping that that after 10seconds, if it is still “hunging”, please stop.
I am using @retry(wait = wait_random_exponential(min =1, max = 60), stop = stop_after_attempt(6)) def chatcompletion_with_backoff(**kwargs): to hope to handle this.
HOWEVER, there are no “timeout”. I have been monitoring the process. There are mulititimes that the the request being going on >minutes, and yet no Timeout is thrown out.
See this for the similar complaints…
I mean, basic coding principle, things should “do what they say, say what they do”… it’s clearly not the case here.
any guidance pls? Just keep “trying” and wait for things to clear, unfortunately is not an option
Exactly, does not give me the confidence to roll my app functionality out to further users at all, unless things get a bit more stable.
I’ve faced the same timeout issue. In the Python API, if you pass the parameter request_timeout to openai.ChatCompletion.create instead of timeout then it will throw a Timeout exception after n seconds. Hope it helps
But yes, I’ve been facing these issues as well…
We have exactly the same problem with timeouts.
thank you for pointing out therequest_timeout
same here! I plan to use claude to see if it is more stable
