v2.8.5

Traces now report latency and error stats grouped by agent, team, workflow, or endpoint, along with tool and model call stats, so you can see exactly where time goes and where things fail rather than reading traces one run at a time. It's available on PostgresDb and SqliteDb, giving you aggregate performance and reliability views straight from the database you already trace to.

View Agno’s Tracing docs for more.

View the platform ops agent cookbook for implementation.