Load Testing Framework for LLM APIs

Build a load testing framework for LLM-powered APIs.

## API Endpoints
{{api_endpoints}}

## Expected Load
{{expected_load}}

## SLA Requirements
{{sla_requirements}}

Implement comprehensive load testing:

```python
class LLMLoadTester:
    def configure_scenario(self, scenario: LoadScenario) -> None:
        """
        Scenario types:
        - Constant load
        - Ramp up/down
        - Spike testing
        - Soak testing
        """
        pass
    
    async def run_test(self, duration: int) -> LoadTestResults:
        """Execute load test with metrics collection"""
        pass
    
    def generate_report(self, results: LoadTestResults) -> str:
        """Generate detailed analysis report"""
        pass
```

Metrics to capture:
- Latency percentiles
- Throughput
- Error rates
- Token usage patterns
- Cost under load