Even the response says I still have a bunch of remaining requests and tokens.
We barely have any volume on this but it’s still failing all the time.
Any idea what’s going on?
Same over here. Not always but from time to time. Also, experiencing weaker output quality
Any news here?
I have already contacted support like 10-12 days ago but also still waiting for a response.
Do you happen to have the error message text? Mine indicates there was some sort of server side error in processing, and it started happening last night some time. Previous to that, it was working fine. I’ve seen other people posting about the same problem, so I suspect there is an issue on the OpenAI side of things.
In the meantime, I have had to put a 60s pause in between retries in order to get successful responses. I suspect OpenAI is struggling to meet demand, and perhaps people are aggressively retrying resulting in further pressure on the infrastructure.
i have the same issue, i was planning for a beta launch of an app but if i can’t even handle more than 1 request/2minutes i don’t know what to do
it’s been this long? have you tried another api key/another account?
Same here. I tried to use another API key, use another model (da-vinci3 & chatgpt) but got the same issue since about two days. Is someone tried to use another account ?
The error message is in the screenshot at the top.
Basically 429, too many requests
Yes.
I’ve seen a bunch of people on Reddit saying they don’t get responses at all so I don’t have high hopes TBH.
We’re trying to work around it for now…
In the provided response, the headers indicate that you are indeed hitting rate limits. The headers 'x-ratelimit-limit-requests': '3500' and 'x-ratelimit-remaining-requests': '3498' suggest that the server allows a maximum of 3500 requests and you have 3498 remaining before hitting the limit.
However, if you believe you are receiving 429 errors without actually hitting these limits, there could be a few reasons for this:
- Server Misconfiguration: It’s possible that the server you’re interacting with has a misconfigured rate limiting system, resulting in erroneous 429 errors. In such cases, reaching out to the server administrator or support team can help resolve the issue.
- Request Bursting: Even if you have not hit the overall rate limit, sending multiple requests within a short period can trigger per-second or per-minute rate limits. Make sure you’re spacing out your requests to stay within the allowed limits.
- IP-based Limits: The rate limits could be imposed on an IP address basis, and if multiple users or applications are sharing the same IP address, the collective requests might exceed the limit. In this case, you can try using a different IP address or contact your network administrator.
- Network or Proxy Issues: Sometimes, intermediate proxies or network infrastructure can introduce rate limits that are not directly related to the server. If you’re behind a proxy, VPN, or firewall, try bypassing them temporarily to see if the issue persists.
If you’ve ruled out these possibilities and the 429 errors persist, please react me out for an help and it will be sincerely done
1 Like
What is a good way to do this? I got hit with this limit error and I’m sure I was no where near 3500 requests per minute. I assume it was Request Bursting as I was processing a batch that was chunking and sending those chunks one after the other.
What is a suggestion for pacing something like that? But a 1 second sleep between each request? What do others do?