[CRITICAL] Persistent Global Saturation - Gemini 3.1 Pro High - Recurring HTTP 503 Errors

Google Antigravity
1

Service: Google Antigravity / Gemini 3.1 Pro High
Issue: Systemic “MODEL_CAPACITY_EXHAUSTED” (HTTP 503)
Duration: 4+ hours of persistent downtime

Summary of the situation:
A severe and prolonged service instability regarding the gemini-3.1-pro-high model is being reported. This is not a transient traffic spike; the error has been recurring for several hours with increasing frequency, making professional workflows impossible to maintain.

Technical Logs:

  • Status: UNAVAILABLE

  • Reason: MODEL_CAPACITY_EXHAUSTED

  • Domain: googleapis.com

  • Example Trace ID: 0xa3188c7f09eaaa15

Observation:
This does not appear to be a user-side quota issue (Tier limits are not reached). The error explicitly points to a physical infrastructure saturation (TPU/GPU clusters) at the server level. Even with high-tier access, the service is rejecting requests systematically.

Expectations:

  1. Can a major incident be confirmed for this specific model/region?

  2. Is there an estimated time for recovery (ETA) for the compute capacity to be restored?

  3. Simple workarounds like switching to “Flash” are not viable for high-precision tasks currently in production.

2

Facing the same issue too the whole afternoon.

3

I am afflicted by this identical predicament; the error persistently manifests, regardless of the hour of its utilization.