Anyone else facing this? Strangely I am able to hit the API fine when in development but it is failing in production.
Yeah, you can keep up to date usually on the Status page for OpenAI.
Partial outage across ChatGPT and the API Subscribe
Monitoring - A fix has been implemented and we’re seeing services back to normal, but we are still monitoring to ensure we’re fully restored.
Nov 07, 2023 - 20:47 PSTUpdate - API and ChatGPT continue to have periodic outages – we have implemented a couple mitigations but are still investigating the issue.
Nov 07, 2023 - 20:33 PSTInvestigating - We are currently investigating this issue.
Nov 07, 2023 - 19:52 PST
For reference, here is the error on my production server:
Error: Unknown error occurred
at (node_modules/openai/error.mjs:49:13)
at (node_modules/openai/core.mjs:263:20)
at (node_modules/openai/core.mjs:303:23)
at (app/api/advise/route.ts:43:21)
at (node_modules/next/dist/esm/server/future/route-modules/app-route/module.js:199:36) {
status: 429,
headers: {
cf-as-number: ‘16509’,
cf-ray: ‘822b6af442e23fd4-SIN’,
connection: ‘keep-alive’,
content-length: ‘2’,
content-type: ‘application/json’,
date: ‘Wed, 08 Nov 2023 05:29:21 GMT’,
retry-after: ‘79476’,
server: ‘cloudflare’,
set-cookie: ‘__cf_bm=c9RwrbY6g8CPx2kyPZcLI1_WOEKKH2ZTiOvQNolK2vg-1699421361-0-AcembHRmdFNGoBbZ0Xf2i+rsuCyD5GCglAxRSVowIRFfVaBahgVNn//gRMcO6+FWApoj2yh2lcRghVYw9VcOGpE=; path=/; expires=Wed, 08-Nov-23 05:59:21 GMT; domain=.api.openai.com; HttpOnly; Secure; SameSite=None’
},
error: undefined,
code: undefined,
param: undefined,
type: undefined
}
We are also getting random 429 errors in production
It is not related to our quota or usage limits since we aren’t anywhere near hitting them
We are using Cloudflare Workers and its only happening to some users. Retrying does not help at all - it seems to either work for a particular user or it doesn’t at all, leading me to believe this is IP based (since Workers runs close to the user so the IP will be based on that)
The response body is simply {}
Same here with @connor.
Always returns a status code of 429. The response body is also empty. It works from an environment other than Cloudflare Workers (for example, OpenAI Playground).
This error occurs even without using the preview model.
The following is a response from the chat completion API:
{
"status": 429,
"headers": {
"cf-as-number": "132892",
"cf-ray": "822c04a3f04fc49b-SEA",
"connection": "keep-alive",
"content-length": "2",
"content-type": "application/json",
"date": "Wed, 08 Nov 2023 07:14:16 GMT",
"retry-after": "74667",
"server": "cloudflare",
"vary": "Accept-Encoding"
}
}
Might or might not be related to the original outage, so I’ll keep everything in this thread. With that said, it sounds like Cloudflare might be affected by something else? If it doesn’t clear up (or it does), you might want to report to help.openai.com too to make sure someone knows what’s happening - if Cloudflare or OpenAI doesn’t pick up on it…
I am seeing the exact same thing on Vercel.
Also the forum does not let you post short replies. Is this really a good idea?
Finally working for me but that was uncomfortably long downtime
I’m still experiencing issues on my end. The chat history in ChatGPT is not loading properly, and every time I click to verify that I’m human through Cloudflare, nothing gets sent afterward.
It appears that there might be a Cloudflare issue on my end.
Still down for me, 429 errors despite not being anywhere near our limits.
Has been occurring now for 4-5 hours. Having a serious impact.
I am still experiencing this - I have no idea what to do now. I’ve tried everything.
Anyone else?
Yes, but the status says the outage has been resolved. I’m confused.
Mine fixed itself around 30 mins ago - I think everyones is getting resolved at different times.
It seems like I’ve been left behind. It’s a caution to everyone heavily relying on AI