Large ChatGPT Project stuck on “Searching project context” with repeated HTTP 429s since Sep 3 incident

[gustavaopere]

Hi,

I’m a ChatGPT Plus user and since the ChatGPT service incident on September 3, 2026, one of my large Projects has become almost unusable.

Before the incident, this Project was working normally.

Actual behavior

• New chats inside the affected Project get stuck on “Searching project context” for several minutes.
• They eventually fail.
• Chrome DevTools shows multiple repeated HTTP 429 responses while the Project is trying to retrieve context.
• I also captured:
  • GET /backend-api/tasks → 500 Internal Server Error
  • POST /backend-api/conversation/init → 404 Not Found
• At some points I also received “Too many requests” when trying to access conversations.

Troubleshooting already performed

• Chrome normal session
• Chrome Incognito
• Cleared all ChatGPT site data
• Tested on mobile
• Tested on another device
• Created a brand-new empty Project → works normally
• Created another new Project and moved all chats from the affected Project into it → the same problem persisted

The last test seems particularly relevant: an empty Project works, but after moving the affected set of chats into a new Project, the context-loading problem follows them.

Normal non-Project chats generally work.

Environment

• ChatGPT Plus
• ChatGPT Web
• Windows 10
• Chrome 152
• Also reproduced on mobile
• Brazil / Brasília timezone (UTC-3)
• Issue started September 3, 2026 and is still occurring September 4

I already opened a case with OpenAI Support and provided the troubleshooting results.

I’m mainly trying to determine whether other users with large/existing Projects started seeing the same behavior after the September 3 incident.

Is anyone else seeing Projects stuck on context retrieval, repeated 429s, or similar failures since yesterday?

[Slavax]

Similar issue, I didn’t dig much, eventually resolved by waiting and delegating more onto Work

It seems you gamble and wait the 25 minute window to see if you get a result, sometimes it’s just a streaming error other times absolutely nothing was done. Not sure if another hidden nerf or something was broken unintentionally due to astra roll out demand

And yeah I’m not trying to cheat chat by running long tasks it’s just rather simpler ones around 5-10 minute ones now take more time

[gustavaopere]

Thanks — that does sound somewhat similar, although in my case the failure is much more persistent.

To clarify my exact behavior:

• The affected Project was the Project I used the most and it worked normally before the Sep 3 incident.
• Since then, it takes a very long time to load/retrieve Project context.
• It rarely manages to produce a response.
• Most attempts stay on “Searching project context” for several minutes and then end in an error.
• This also happens in completely new chats created inside the affected Project.
• I’m using GPT-5.6 Sol with High intelligence/reasoning.
• A brand-new empty Project works normally.
• I created another new Project and moved all of the affected chats into it, and the same problem followed the chats/context.

So for me this is not just that 5–10 minute tasks became slower. Even a simple “hi” may spend several minutes retrieving Project context and then fail.

During reproduction I also captured repeated HTTP 429 responses, plus backend 500/404 errors in DevTools.

Did your affected Project als

o have a large amount of previous chat history/context, and did simple new messages inside it sometimes fail too?