A real-time sports data pipeline that runs unattended
Gaureesh Vats Shukla
SportsForge: a real-time sports data pipeline that runs unattended
The goal
Track live events for hours at a time without anyone watching, and keep going when the data feed breaks.
What I built
5 services in production, polling every 7 seconds, with the interval set from measured feed latency
State saved per event, so a restart resumes where it stopped instead of replaying
An override map for when the provider renames teams or players
Output held until delivery is confirmed, so no alert goes out twice
Containerised, deployable on the client's own server
The result
Events of 4 hours run end to end on their own. When the feed drops mid-event, the system picks up from saved state and carries on.
What runs in production
What happens when the feed breaks.
Like this project
Posted Oct 10, 2026
Real-time sports data pipeline: 5 services in production, polling every 7s. When the feed breaks mid-event, it resumes from saved state.
