Error Tracking

Track failed API calls to monitor error rates and identify issues.

Why Track Errors?

• Monitor success rates in your dashboard
• Identify problematic prompts or models
• Get alerted when error rates spike
• Debug issues faster with error logs

Basic Error Tracking

import { CostLens } from 'costlens';

const costlens = new CostLens({ 
  apiKey: process.env.COSTLENS_API_KEY 
});

const start = Date.now();
try {
  const result = await openai.chat.completions.create(params);
  await costlens.trackOpenAI(params, result, Date.now() - start);
} catch (error) {
  // Track the error
  await costlens.trackError(
    'openai',
    params.model,
    JSON.stringify(params.messages),
    error,
    Date.now() - start
  );
  throw error; // Re-throw to handle in your app
}

What Gets Tracked

When you track an error, we capture:

• Model - Which model failed
• Input - What was sent to the API
• Error message - The error details
• Latency - How long before it failed
• Timestamp - When it happened

Common Error Types

Error: Rate limit exceeded
Solution: Implement exponential backoff or upgrade your API plan

Error: Invalid model specified
Solution: Check model name spelling and availability

Error: Invalid API key
Solution: Verify your API key is correct and active

Viewing Errors in Dashboard

Go to your dashboard to see:

• Overall success rate percentage
• Error rate trends over time
• Most common error messages
• Which prompts have highest error rates

Best Practices

1. Always use try/catch - Never skip error handling
2. Track before re-throwing - Log to CostLens then handle in your app
3. Include context - Use promptId to identify problematic prompts
4. Set error rate alerts - Get notified when errors spike

Error Rate Alerts

Set up alerts in Settings to get notified when:

• Error rate exceeds 5%
• Specific prompt has high failure rate
• New error types appear

Next Steps