Enhancing Netflix Reliability with Service-Level Prioritized Load Shedding
https://www.infoq.com/news/2024/11/netflix-load-shedding/
Method
"services can determine when to shed traffic from a particular bucket based on CPU utilization"
Population
none stated
What it does not show
Does not quantify how often the limiter over-sheds legitimate high-priority traffic. A single incident rather than a study across outages, and no comparison against alternative overload controls such as autoscaling alone.
Anirudh Mendiratta (Netflix), reported by InfoQ
A concurrency limiter that categorises requests by criticality held user-initiated request availability above 99.4% during a real outage while shedding prefetch traffic to 20% availability.
Tier IV: Single-organisation engineering report of one production overload incident with concrete availability numbers; no control group. Accessed via a secondary source because the primary blog returned 403, which is disclosed rather than hidden.