Report full API response on failure for ease of debugging #2262
Labels
clarification
provide explanation for goose's behavior
Comments
|
It turns out the rate limit was due to #2263 |
|
yes, previously we logged all info to the screen but sometime the payload may be very long and confusion, so we only show the error message in the screen and log all payloads to the log. |
|
Where does the goose log go by default? |
|
On OSX it appears to be ~/Library/Application Support/Goose/logs/main.log |
|
And more in ~/.local/state/goose/ ? |
|
yes, you can find more mapping in #1153 and https://block.github.io/goose/docs/guides/logs |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
So, I understand the standard guidance when hitting rate limit with LLM providers is to use OpenRouter. That being said, I have a paid API key from Google and I don't think I should be rate limited, and I'm interested in diagnosing why. Right now, all the info I get is:
https://www.googlecloudcommunity.com/gc/AI-ML/Assistance-Needed-Quota-Limit-Increase-for-Generative-Language/m-p/815976?utm_source=chatgpt.com implies that I should have gotten detailed information from the API response about what exactly was the rate limit triggered. It would be supremely helpful if this was surfaced by default.
I'm also not sure if there's a way to get at the raw API logs from the UI, I see https://book.goose.rs/logging/debug.html?utm_source=chatgpt.com but I tried:and it didn't seem to do anything.Oops apparently this is a completely different goose LOLThe text was updated successfully, but these errors were encountered: