On June 24, 2026, Spotify's video transcoding infrastructure hit maximum capacity, causing multi-hour delays in podcast episode publication. Five converging factors caused the outage: insufficient headroom for traffic spikes, a concurrent batch reprocessing job, recent quality improvements that increased per-item processing cost, a prioritization system malfunction that treated batch jobs equally to new submissions, and a resource scheduling bug underutilizing compute by ~10%. The team stopped the batch job, deployed a scheduling fix, and brought additional capacity online overnight. All queues cleared by 01:02 UTC on June 25. Follow-up actions include a 67% capacity increase, fixed resource scheduling, earlier monitoring alerts, improved prioritization, and better backpressure mechanisms throughout the pipeline.
