fix(anthropic): incorrect exception raised when context limit exceeded

[yoonian]

Description

Added handling for "input length and max_tokens exceed context limit" to raise ContextWindowOverflowException

Related Issues

Documentation PR

Previously, a generic error was raised instead of ContextWindowOverflowException.
Now it correctly maps the "input length and max_tokens exceed context limit" error.

Handle Anthropic/LLM API errors containing "input length and max_tokens exceed context limit" by explicitly raising ContextWindowOverflowException instead of generic RemoteProtocolError(for strands-agent 0.12.0) or EventLoopException or anthropic.BadRequestError(for anthropic 0.71.0, strands-agent 0.13.0)

Type of Change

Bug fix

Testing

• I ran hatch run prepare

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• I have updated the documentation accordingly
• I have added an appropriate example to the documentation to outline the feature, or no new docs are needed
• My changes generate no new warnings
• Any dependent changes have been merged and published

---

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[dbschmigelski]

Under what conditions where you seeing this?

When I ran a test using

def test_input_and_max_tokens_exceed_context_limit():
    """Test that triggers 'input length and max_tokens exceed context limit' error."""
    # Create a model with very high max_tokens
    model = AnthropicModel(
        model_id="claude-3-7-sonnet-20250219",
        max_tokens=200000,  # Very high max_tokens
    )
    
    # Create very large input that when combined with max_tokens exceeds context window
    large_text = "This is a very long text. " * 50000  # ~150k tokens
    
    agent = Agent(model=model)
    
    with pytest.raises(ContextWindowOverflowException):
        agent(large_text)

I ended up seeing

{
    'type': 'error',
    'error': {
        'type': 'invalid_request_error',
        'message': 'max_tokens: 200000 > 64000, which is the maximum allowed number of output tokens for claude-3-7-sonnet-20250219'
    },
    'request_id': 'req_XXXXXXXX'
}

Which we will have to see if we should handle this or not but I want to reproduce the message you are proposing.

[yoonian]

I am using the model 'claude-sonnet-4-20250514'

[zastrowm]

I also was not able to reproduce the:

input length and max_tokens exceed context limit

case. However, in attempting to reproduce I did get a "prompt is too long" error, which I created #1137 to address.

@yoonian do you have code that reproduces the error that you saw?