I am receiving a 429 Too Many Requests error for the Vertex AI API and my application has stopped working. The error message indicates that a quota has been exceeded, but I cannot find this quota in the Google Cloud Console to request an increase.
The full error is: Quota exceeded for aiplatform.googleapis.com/generate_content_requests_per_minute_per_project_per_base_model with base model: imagenegeneration. Please submit a quota increase request: https://cloud.google.com/vertex-ai/docs/generative-ai/quotas-genai.","status":"RESOURCE_EXHAUSTED"}}
I have tried the following without success:
Searching the Quotas page in the IAM & Admin section.
Filtering by the full quota name and by “imagenegeneration.”
The quota does not appear to be adjustable in the console.
I am unable to create a support case due to my account’s support plan.
Could a Googler please assist me by creating an internal request to increase this quota?
I invite you to keep an eye on the Vertex AI API/Service Details page. In your case, you should filter by Metric: aiplatform.googleapis.com/generate_content_requests_per_minute_per_project_per_base_model.
imagenegeneration as Base Model seems to be a generic name. To find the corresponding Model, you may have to refresh the page while you’re using your app, and the API nearing its usage limits might then appear.
Also, Logs Explorer is a good way to troubleshoot quota-related issues.
I see that you’re using Firebase, so you may be interested in looking at Rate limits and quotas from Firebase with Vertex AI.
Last but not least, if possible, distributing the load over multiple regions is an effective way to avoid reaching quota limit too soon.
In addition to @LeoK response, you’re encountering a 429 “Too Many Requests” error because your application has exceeded the per-minute quota for the imagenegeneration base model in Vertex AI. To resolve this, you can implement exponential backoff in your application to reduce request spikes, consider switching to the global endpoint for better quota distribution, or explore subscribing to Provisioned Throughput for guaranteed capacity.
In your case, you can check and follow the steps in view and edit quotas, noting that the request will be subject to review.
Thank you for the replies. I was able to send a request for a quota increase and it was immediately denied with no reason given and provided a link to live chat support. Which was only for billing and not tech support. Do I really need to PAY for support, just to request a quota increase?
