Every time I try to talk, you are asked to limit my usage

[sdr0x07b6]

Type: Bug

"Sorry, your request has been rate limited. Please wait 84 seconds and try again."

The "1 second" displayed in the message is not exactly 1 second, but I have to wait about 5 times the amount of time I am asked to wait before I can talk.

I am using it in Japanese, so the exact message that appears is as follows:

「申し訳ございません。要求はレート制限されました。84 seconds を待ってから、もう一度お試しください。」

Extension version: 0.21.0 VS Code version: Code 1.94.0 (Universal) (d78a74bcdfad14d5d3b1b782f87255d802b57511, 2024-10-02T13:08:12.626Z) OS version: Darwin arm64 23.6.0 Modes:

System Info

|Item|Value| |---|---| |CPUs|Apple M2 (8 x 2400)| |GPU Status|2d_canvas: enabled
canvas_oop_rasterization: enabled_on
direct_rendering_display_compositor: disabled_off_ok
gpu_compositing: enabled
multiple_raster_threads: enabled_on
opengl: enabled_on
rasterization: enabled
raw_draw: disabled_off_ok
skia_graphite: disabled_off
video_decode: enabled
video_encode: enabled
webgl: enabled
webgl2: enabled
webgpu: enabled
webnn: disabled_off| |Load (avg)|3, 4, 4| |Memory (System)|24.00GB (0.07GB free)| |Process Argv|--crash-reporter-id dce19cec-7600-4b8a-8588-e3edc362db29| |Screen Reader|no| |VM|0%|

A/B Experiments

``` vsliv368:30146709 vspor879:30202332 vspor708:30202333 vspor363:30204092 vscod805:30301674 binariesv615:30325510 vsaa593:30376534 py29gd2263:31024239 c4g48928:30535728 azure-dev_surveyone:30548225 2i9eh265:30646982 962ge761:30959799 pythongtdpath:30769146 pythonnoceb:30805159 asynctok:30898717 pythonmypyd1:30879173 2e7ec940:31000449 pythontbext0:30879054 dsvsc016:30899300 dsvsc017:30899301 dsvsc018:30899302 cppperfnew:31000557 dsvsc020:30976470 pythonait:31006305 dsvsc021:30996838 g316j359:31013175 a69g1124:31058053 dvdeprecation:31068756 dwnewjupytercf:31046870 impr_priority:31102340 nativerepl1:31139838 refactort:31108082 pythonrstrctxt:31112756 flightc:31134773 wkspc-onlycs-t:31132770 nativeloc1:31134641 wkspc-ranged-t:31151552 cf971741:31144450 defaultse:31146405 iacca1:31150322 notype1cf:31151524 5fd0e150:31146321 ```

[lramos15]

The seconds being incorrect is a known issue. The rate limit message is just due to using a large amount of tokens within copilot chat which puts you in the top 0.01% of users. We are working to adjust the rate limit as to not affect heavy users of Copilot

[lramos15]

Please give VS Code insiders + pre-release a try. We've done some token optimizations there that should make it less likely for you to hit the rate limit.

[sdr0x07b6]

If it is long, you are told to wait 15 minutes. This means I actually have to wait an hour or more before I can use it. Not sure what you mean by 0.01%, but I don't use it that excessively.

I think I will forgo VS Code insiders + pre-release. Github Copilot has always seemed very unstable, even in stable, and Github Copilot has always been very stressful for me. I am not interested in installing a development version that may be even more unstable.

When do you plan to release a fix for this glitch? I don't mean to overuse Github Copilot, but I keep getting this message almost every day and it is unusable.

This is a problem not only for the productivity of my work, but also for the money I have already paid.

[sdr0x07b6]

It's been a month since the other day, and now the problem has reoccurred.

申し訳ございません。要求はレート制限されました。2 hours 27 minutes を待ってから、もう一度お試しください。

They say you have to wait 2 hours 27 minutes. In reality, it may be five times longer. This is not working. Hasn't this been improved? Is it your goal that eventually there will be no such messages at all?

[lramos15]

Hasn't this been improved?

This has been improved we have tripled the limit. The number of users impacted by this are minimal. If you have long history I would recommend starting new conversation threads as well to avoid sending possibly irrelevant excess tokens. You are using a significant amount of tokens and there is no real bug here.

[sdr0x07b6]

Thank you for your advice to use threads creatively. I will try to do so.

However, the way I use chat is not to have the AI ​​answer the questions one by one, but to continue the conversation and thinking together with me. In other words, threads always have a "history" or "context". I think many users use it like that.

If you move to a new thread when the thread gets long, you can avoid the restriction, but each time you move to a new thread, you will have to explain the "history" and "context" that you have been talking about up until that point from scratch again, so it is difficult to move to a new thread.

GitHub Copilot Chat is an AI specialized in the programming field, so there should be more use cases like this than other AI services. It would be great if GitHub Copilot Chat itself had a function to "create a new chat while moving the context".