It seems that Google uses 429 to mean “our server is too busy right now”, instead of the standard 503 Service Unavailable: Standard PayGo | Generative AI on Vertex AI | Google Cloud Documentation
If you receive a
429error, it doesn’t indicate that you’ve hit a fixed quota. It indicates temporary high contention for a specific shared resource.
In my opinion it is quite rude to return 429 in this instance, implying that it’s something that the user caused and has the power to fix. In fact, it may have nothing to do with what the user requested, and everything to do with Google just not having enough resources to service all its customers at the same time.
A cynic might assume that returning a 4xx error instead of a 5xx error is done to fudge the stats on uptime and availability… 