Trajectory ID: bc4b2caa-a625-44e2-b2e6-84b******* Error: HTTP 503 Service Unavailable Sherlog: TraceID: 0xe8fd44ff******** Headers: {"Alt-Svc":["h3=\":443\"; ma=2592000,h3-29=\":443\"; ma=2592000"],"Content-Length":["453"],"Content-Type":["text/event-s

Trajectory ID: bc4b2caa-a625-44e2-b2e6-84b*******
Error: HTTP 503 Service Unavailable
Sherlog:
TraceID: 0xe8fd44ff******
Headers: {“Alt-Svc”:[“h3=”:443"; ma=2592000,h3-29=“:443”; ma=2592000"],“Content-Length”:[“453”],“Content-Type”:[“text/event-stream”],“Date”:[“Thu, 17 Sep 2026 03:38:05 GMT”],“Server”:[“ESF”],“Server-Timing”:[“gfet4t7; dur=621”],“Vary”:[“Origin”,“X-Origin”,“Referer”],“X-Cloudaicompanion-Trace-Id”:[“e8fd44ff******”],“X-Content-Type-Options”:[“nosniff”],“X-Frame-Options”:[“SAMEORIGIN”],“X-Xss-Protection”:[“0”]}

{
“error”: {
“code”: 503,
“details”: [
{
“@type”: “type.googleapis.com/google.rpc.ErrorInfo”,
“domain”: “cloudcode-pa.googleapis.com”,
“metadata”: {
“error_number”: “2010”,
“model”: “gpt-oss-120b-medium”
},
“reason”: “MODEL_CAPACITY_EXHAUSTED”
}
],
“message”: “No capacity available for model gpt-oss-120b-medium on the server”,
“status”: “UNAVAILABLE”
}
}

Hi @zheng_nie, Welcome to the Forum,

Your account and setup are completely fine—error 503 with MODEL_CAPACITY_EXHAUSTED (Error 2010) is a server-side capacity spike on Google’s backend specifically for the gpt-oss-120b-medium model cluster on cloudcode-pa.googleapis.com.

So the quick way around this:

  1. In the bottom model selector of your chat canvas, switch to Gemini 3.8 Flash or Gemini 3.5 Pro. These run on Google’s primary TPU infrastructure with much higher capacity and will work immediately.
  2. Start a fresh conversation (Cmd/Ctrl + N) after switching models, as the current trajectory terminated on the failed 503 response.
  3. If you need gpt-oss-120b-medium specifically then wait 3–5 minutes for the inference queue to drain and retry.

Please switch to Gemini 3.8 Flash in the model dropdown and let us know is your agent resumes working! or the issue still persists.