I've integrated gemini api into a project of mine and sending prompts with images.
I am on the free tier, and this is the usage I see on your console:
As you see I never reached the quota limit, problem is I am getting 429 as a response most of the time (these are single calls, note that are well below the free tier quota limit):
warning | 2025-02-12 17:34:30 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 17:34:44 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 17:36:06 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 17:36:23 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 17:37:17 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 17:37:22 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 17:43:01 | frigate.genai.gemini | Gemini returned an error: 503 GOAWAY received; Error code: 0; Debug Text: session_timed_out
warning | 2025-02-12 17:43:05 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:01:04 | frigate.genai.gemini | Gemini returned an error: 503 GOAWAY received; Error code: 0; Debug Text: session_timed_out
warning | 2025-02-12 18:24:32 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:26:18 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:26:19 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:26:29 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:26:50 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:36:09 | frigate.genai.gemini | Gemini returned an error: 503 GOAWAY received; Error code: 0; Debug Text: session_timed_out
warning | 2025-02-12 18:46:11 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:46:54 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:47:07 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:47:50 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:48:10 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:51:15 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:51:50 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:51:51 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:51:56 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:55:09 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:55:12 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
warning | 2025-02-12 18:55:12 | frigate.genai.gemini | Gemini returned an error: 429 Resource has been exhausted (e.g. check quota).
Out of ~20 requests over a few hours, only 1 succeded, all the others were 429.
I reached out to the maintainers of the program is doing the calls, and they suggested reaching out to you, for reference: https://github.com/blakeblackshear/frigate/discussions/16526
An ever better view: over the last 6 hours there have been 71 requests, 84.51% of which were 429, even though the requests were well spread over 6 hours (so never reaching the 15 req/min free tier cap).
What's wrong?
Thanks.
This is happening to me as well.
This view shows probably better that I am not reaching any of the quotas:
even though I am getting basically only 429s:
I am facing the same issues as well.
I am using the free plan with an average of 1.5 requests per minute and under 1500 tokens per request and I run into quota exceeded errors "google.api_core.exceptions.ResourceExhausted: 429 Resource has been exhausted (e.g. check quota)."
Same here even with a paid plan
If you create a new api it works once, then doesn't work after
Happening to me on a paid account when I try to use grounding via tools (google_search). It had been working fine all day and then a solid wall of 429s. Checked my quotas and nothing is even close.
Tried to adjust my quota just for funzies and I can only adjust it down. Super helpful Google!
Figured there must be some unpublished limit and then found the below on the dev pricing page. I am on the paid tier so it's beyond me why I can't make more grounding calls and just pay for them. To be honest my experience with the Google AI apis has been a bit of mess.
Facing the same issue on paid tier. I can run 1500 requests with ground, but instead of switching to paid after the initial 1500 requests, I just keep getting the 429, until next day.
This is on Flash 2.0. Flash 1.5 works and charges me, but I seem to recall it had the same problem when it was 1500 free / rest paid as well.
Would love to know if there is a fix, since we are production ready and need to scale soon.
User | Count |
---|---|
2 | |
2 | |
1 | |
1 | |
1 |