Logfire AI — Automate Logfire with Runner
Ask what's happening in production and get a straight answer fast.
Runner uses Logfire through a native MCP server for OpenTelemetry trace search, span analysis, and SQL-based observability queries.
What Runner does in Logfire
- Search spans by operation name, error status, and time range.
- Fetch full trace chains to understand distributed call flows.
- Use SQL queries for aggregations like error rates and latency percentiles.
Example asks
Search for error spans
Find failed spans in the auth service from the last 2 hours.
Production issues surface in the workflow without switching to the observability dashboard.
Trace a specific request
Pull the full trace for this slow request and show me where the latency is.
Runner maps the call chain so you can see where things broke down.
Query error rates with SQL
Run a SQL query to show the error rate by endpoint over the last 24 hours.
Custom observability queries run inside the workflow instead of requiring a dashboard switch.