Platform Analytics

Know which MCP servers are actually earning their keep

Fleet telemetry lives across several tables, so every question about adoption, errors, or dormant connections becomes a manual join. These plans make that snapshot repeatable and schedulable.

Book a demoBrowse every plan

What breaks when you do this by hand

  • Understanding how one MCP server is really performing means joining several telemetry tables by hand, every time.

  • There's no quick, repeatable snapshot of which servers drive usage.

  • Errors and dormant connections stay invisible until someone goes looking.

What you get instead

  • A fleet-wide view of usage, errors, and dormant connections

  • Per-server performance without hand-written joins

  • Reporting that runs on a schedule instead of on request

This is the whole thing

The opening of MCP Server Insights — the system prompt, its typed parameters, and the tools it's allowed to reach, all declared up front. No canvas, no hidden nodes. 207 lines of source you can review in a pull request.

Read all 207 lines →

mcp-server-insights.fml

Copy
system("You are an analytical assistant querying BigQuery for MCP server telemetry.")

parameter("servername", type=string, title="Server Name")

require mcp BigQuery

components {
    schema("TenantStats") {
        tenant_name: string
        tenant_id: string
        status: string
        total_tool_calls: int
        total_errors: int
        error_rate: float
        active_users: int
        distinct_tools_called: int
        active_days: int
        first_call: string
        last_call: string
        days_since_last_call: int
        connected_but_never_called: bool
    }

2 plans you can run today

Every one is a typed FML plan. Open it, read the source, and run it — nothing here is a mockup.

MCP Server Insights

Understanding how one MCP server is really performing across the fleet means joining several telemetry tables by hand every time.

Platform PMs
DevOps
Engineering leaders

Read the plan

MCP Fleet Overview

There's no quick, repeatable snapshot of which MCP servers are actually driving usage and where errors and dormant connections concentrate.

Platform PMs
DevOps
Leadership

Read the plan

What every plan here guarantees

Zero prompt drift

Every plan is a versioned contract. Run 1 and run 10,000 behave identically.

Scoped sessions

Each LLM call sees only the context it needs — no one giant prompt, no context rot.

Typed output

Plans return validated objects pinned to a schema, not text you have to parse.

Reusable like an API

Parameterise once and call it from anywhere — versioned, auditable, shareable.