Case Study: Peloton achieves scalable real-time streaming and faster incident resolution with Loggly

Streaming spin classes to thousands with Loggly and AWS

Peloton Cycle, founded in 2012, delivers live and on‑demand indoor studio cycling classes to thousands of riders at home, with an average of 350 people per class and metrics updated every second—resulting in some 17 billion API calls per day. A key challenge was making on‑demand classes feel as engaging as live sessions: the competitive leaderboard must combine live and historical data, apply accurate participant filtering, and retain permanent metrics while engineers manage logs and application behavior across a ~300‑machine virtual production cluster.

To solve this, Peloton runs on AWS and uses Loggly to collect and analyze petabytes of log data, automatically backing logs to Amazon S3, eliminating manual log rotation and storage hardware, and providing cross‑system visibility. Loggly’s insights speed diagnosis and reduce mean‑time‑to‑repair so engineers can release features faster and focus on product innovation: “The type of fast diagnosis, to find out if issues are one-off or systemic, is something that we haven’t found outside of Loggly.” — Bryan Tinsley, Site Reliability Engineer, Peloton

Peloton

Bryan Tinsley

Site Reliability Engineer