Hi everyone,
Hope everyone’s doing well. I’m currently reaching out to the community and API engineers because I’ve been dealing with a stubborn 429 Resource Exhausted error for over a dag, and I’m trying to determine if this is a known systemic hangover from recent server outages or a localized quota sync bug.
Context & Setup:
I am a Google AI Pro (2TB) subscriber. My current workflow is strictly CLI-based. I have access to Antigravity and Code Assist, but I have not used Antigravity at all since the last banwave problem during late february and Early March—my setup is purely making calls via the CLI.
The Timeline of the Issue:
Yesterday Daytime: While running some standard CLI tests, I started hitting a general 429 Exhausted error mid-day & afternoon. Assuming this was related to the broader server turbulence happening at the time, I paused my work entirely to avoid spamming the endpoint.
Yesterday Evening: The servers seemed to stabilize. I had a clear window of over an hour where my CLI requests routed perfectly with zero errors.
Yesterday Night to Present (20+ Hours Later): The errors suddenly returned, but the message changed specifically to: PROXY ERROR 429: {“error”: {“message”: “Resource has been exhausted (e.g. check quota).”, “type”: “api_error”, “code”: 429}}.
There’s this new (e.g. check quota) added to the error. I’m highly assured that i didn’t run out of my RPD, TPM or RPM at all as i double checked.
Troubleshooting Done:
I have thoroughly audited my local environment. I have force-killed any sort of Node processes, cleared any tunneling software if i have open, and ensured there are absolutely zero background retry loops or rogue scripts secretly hammering the endpoint. My environment is completely clean, yet even a single manual text prompt instantly returns this specific quota error.
Given that it has been well over 20 hours since the initial 429s, and past the standard Pacific Time midnight reset, my daily quota should have refreshed by now.
My Questions:
-
Is it possible that this is getting “stuck” in a throttled/exhausted state following yesterday’s server turbulence, preventing the daily quota from resetting properly?
-
Is anyone else in the community experiencing this prolonged lockout despite having a completely clean, zero-loop setup?
Any insights from the community or the Google team would be hugely appreciated. I’m completely locked out of using any models at the moment.
Thanks,
Iftehaj (iftx_j.n)
