> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mcp-1.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Request logs (traces)

> A searchable log of every call that has passed through the aggregated endpoint.

Every tool call routed through mcp1's aggregated endpoint is recorded and shown in **Request Logs**. Each entry captures what was called, whether it succeeded, and how long it took.

## Why it exists

Because every AI client call to any backend now flows through a single endpoint, traces give you one place to answer "what just happened" instead of checking logs on each backend individually.

## Filtering and search

Request logs are searchable and filterable — narrow by backend server, tool, or outcome to find a specific call without scrolling through the full history.

## Relationship to health

Traces show what already happened; [Health](/essentials/health) shows current reachability. A backend can be healthy but still show failed calls in traces if, for example, a tool call was rejected for invalid arguments.
