How toLog your application

Track errors in traces

Learn how to effectively track and log errors from LLM results and Tool calls in your AI application traces to improve performance and reliability.

Track LLM errors in your workflow

// Create generation object
const generation = trace.generation({
    id: "generation-id",
    name: "customer-support--gather-information",
    // Additional fields
});
 
// Track error
generation.error({
    message: "Rate limit exceeded. Please try again later.",
    type: "RateLimitError",
    code: "429",
});

Learn how to track complete LLM flows in the LLM logging guide.

Track errors from tool calls

const traceToolCall = trace.toolCall({
    id: "tool-call-id",
    name: "tool-call-name",
});
 
traceToolCall.error({
    message: "Service is currently unavailable. Please try again later.",
    type: "ServiceUnavailableError",
    code: "503",
});

Explore more on tool call tracking in the Tool calls logging guide.

Send feedback for AI application traces

Track token usage and costs

Track errors in traces

Track LLM errors in your workflow

Track errors from tool calls

On this page