Gemini pro models not response half of the requests

Gemini API
1

I am on tier1 plan. But api requests just keep going and never response half of the time. it crush the production. This problem sometimes occur. Is this a known issue? I guess it is issue about server loads

2

Hey! If you are getting 503’s, then yes it is a transient server overloaded. These usually last for short spikes so exponential backoff usually helps or falling back to 3.0 Flash.

3

actually not 503, just never response. Waiting more then 15+ mins for 2k token question still no answer.

It is also happening in 2.5 pro too in Europa day time. So how will we use the google products in our projects?

4

Checking in on the default timeouts, you can set these manually via code like:

```

result = client.models.generate_content(
model=MODEL_ID,
contents=“List about 10 cookie recipes, grade them based on popularity”,
config=types.GenerateContentConfig(
response_mime_type=“application/json”,
response_schema=list[Recipe],
http_options=types.HttpOptions(
timeout=60000 # Pass timeout as a keyword argument to HttpOptions constructor
)
),
)

```

5

so you just suggest the seeting the timeout and retry?