pieces-app / support

11 stars 1 forks source link

Copilot Bug: 'I'm sorry. Something went wrong with processing. Please wait a few minutes and try again.' #193

Open mark-at-pieces opened 2 months ago

mark-at-pieces commented 2 months ago

Software

Desktop Application, Pieces OS, Web Extension, VS Code, JetBrains, Obsidian, Visual Studio, JupyterLab

Operating System

macOS/Window/Linux

Your Pieces OS Version

9.0.2

Early Access Program

Kindly describe the bug and include as much detail as possible on what you were doing so we can reproduce the bug.

Occasionally when using a cloud or local LLM in a Copilot conversation.

One will get this error

'I'm sorry. Something went wrong with processing. Please wait a few minutes and try again.'

Solution:

  1. hard reset the copilot in the copilot setting and retry via "Hard-reset active runtime"
Screenshot 2024-05-03 at 11 23 50 AM

** if you are still having issue, please add a comment to this ticket

mark-at-pieces commented 2 months ago

**Note: this is for documentation of this bug, but please feel free to comment in here if you are still having issues after the hard reset

mark-at-pieces commented 2 months ago

Worst case: if everything is still not working after the hard reset, if you close and reopen PiecesOS then this will also resolve this issue

dotBits commented 1 month ago

To me active runtime hard reset, restart both Pieces OS and Pieces for Developers and even restarting the whole system doesn't make it back to work

philwatt commented 1 month ago

To me active runtime hard reset, restart both Pieces OS and Pieces for Developers and even restarting the whole system doesn't make it back to work

Same here. Have done the hard reset and restarted my Mac. Same error, including after waiting.

I notice that the issue hasn't been updated after @dotBits commented (2 weeks...). @mark-at-pieces What is the fix where the above doesn't work? Uninstall / reinstall?

philwatt commented 1 month ago

To me active runtime hard reset, restart both Pieces OS and Pieces for Developers and even restarting the whole system doesn't make it back to work

Same here. Have done the hard reset and restarted my Mac. Same error, including after waiting.

I notice that the issue hasn't been updated after @dotBits commented (2 weeks...). @mark-at-pieces What is the fix where the above doesn't work? Uninstall / reinstall?

I have tried uninstalling and reinstalling Pieces OS and Pieces for Developers. The deletion process I followed is:

I have a 2019 MacBook Pro with Core i9 processor, 8GB Radeon Pro, 64GB RAM Not sure where to go from here.

JZDesign commented 1 month ago

I'm still getting this issue on a M2 Pro. I've tried the hard reset and restarting Pieces

mark-at-pieces commented 1 month ago

To me active runtime hard reset, restart both Pieces OS and Pieces for Developers and even restarting the whole system doesn't make it back to work

so sorry that i didnt see this, would you mind providing me with the LLm that you were using?

mark-at-pieces commented 1 month ago

To me active runtime hard reset, restart both Pieces OS and Pieces for Developers and even restarting the whole system doesn't make it back to work

Same here. Have done the hard reset and restarted my Mac. Same error, including after waiting.

I notice that the issue hasn't been updated after @dotBits commented (2 weeks...). @mark-at-pieces What is the fix where the above doesn't work? Uninstall / reinstall?

Very interesting behavior, you have done a lot of good stuff here that should 100% have fixed this issue,

note: since you are using the new gpt-4o model

I'm thinking this might be related to different bit of technology within this flow

I'm chatting with a user later this afternoon, but would appreicate a bit of time with you to see if you are having a similar issue as well

https://calendar.app.google/jFcnQQt3EoJMG2kr7

mark-at-pieces commented 1 month ago

I'm still getting this issue on a M2 Pro. I've tried the hard reset and restarting Pieces

I believe this is similar to @philwatt's issue

Thanks for reaching out will be getting this resolved and I will update this thread

philwatt commented 1 month ago

To me active runtime hard reset, restart both Pieces OS and Pieces for Developers and even restarting the whole system doesn't make it back to work

Same here. Have done the hard reset and restarted my Mac. Same error, including after waiting.

I notice that the issue hasn't been updated after @dotBits commented (2 weeks...). @mark-at-pieces What is the fix where the above doesn't work? Uninstall / reinstall?

Very interesting behavior, you have done a lot of good stuff here that should 100% have fixed this issue,

note: since you are using the new gpt-4o model

I'm thinking this might be related to different bit of technology within this flow

I'm chatting with a user later this afternoon, but would appreicate a bit of time with you to see if you are having a similar issue as well

https://calendar.app.google/jFcnQQt3EoJMG2kr7

Hey Mark,

I'm in Melbourne and the time zones are tricky to meet it seems. I think you're right about it being something to do with the model.

It all went wrong after the second time I changed the model. I was using GPT4, then switched to Mistral 7B GPU, and both of these worked brilliantly. I then changed to GPT4, and it all fell apart then.

Next I did the above uninstall/reinstall business, and tried both GPT4 and GPT4o. But after that I tried Mistral 7B GPU again with the same results.

Hope this helps.

Cheers,

Phil

bodzebod commented 2 weeks ago

Same here. I'm running Pieces OS 9.0.6, and Pieces for Developers 2.12.3, on Windows 11 pro. I tried to switch between GPT-4o and GPT-35-Turbo with hard reset, with no change. It used to work last week however. Any log we could inspect for errors?


UPDATE: I tried to use copilot chat in visual studio code, and got the following response:

Sorry, something went wrong with that request: [429] OpenAI: Rate limit or quota exceeded Get Help

Then I retried once more and it unblocked the chat. Now I can get responses from GPT-4o again in both vscode and pieces for developers app.


UPDATE2: I tried to use copilot in IntelliJ IDEA. And got the same error with both GPT4 and GPT3.5 Turbo:

Sorry, something went wrong with that request: [429] OpenAI: Rate limit or quota exceeded

However, when trying to switch to another model, it doesn't show GPT-4o. Instead, I only have: GPT3.5 Turbo, 3 occurences of GPT4 (?!?) and GPT3.5 16k. In addition, Pieces for developers application is no longer able to response, and raises the error again. Same for vscode as well.

bodzebod commented 1 week ago

Just adding my comment about this issue. In my case, I found that my quota on my openai account was fully consumed, and I just needed to add some more credits to my balance. So no issue on Pieces side actually.

mark-at-pieces commented 1 week ago

Most likely some additional error handling that we can do on ourside to inform you as a user, that the quota is used up thank you for adding the issue