
Engineering Principles for Scale

Live streaming at world-record scale with Ashutosh Agrawal (ex-Jio / Disney+ Hotstar)
Topic 10 of 11
Published Feb 12, 2025 · Clip 54:45–59:50 (5:05)
The guest shares his core principles for building large-scale systems: assume anything that can fail will fail (Murphy's Law), which forces scrutiny of every configuration - like database connection pool sizes that cap how far compute can scale before choking the database. He stresses deep, detailed measurement, warning that naive response-time measurement at the processing function misses time spent sitting in load balancer queues. Finally, he says he never uses APMs because they make engineers lazy; without them you're forced to measure every aspect of your code and understand every corner of the system.