arrow_back Observability and Control Theory
Scaling hotstar.com for 25 million concurrent viewers
Submitted by Gaurav Kamboj (@gauravkamboj) on Friday, 26 July 2019
Section: Full talk (40 mins) Category: SRE Section: Full talk (40 mins) Category: SRE
The talk will focus on why traditional autoscaling doesn’t work for Hotstar (Disney’s OTT streaming service) who recently created a global record for live streaming to 25.3 million concurrent viewers. It will talk about challenges in scaling infrastructure for millions and how to overcome it, how we run gamedays before actual games, how we use load testing monster called Hulk to prepare for peak traffic and also how we add chaos (engineering) in between all these epic scale madness.
Why traditional autoscaling doesn’t work for Hotstar
Challenges in scaling
Gamedays and load testing for millions
How we prepare for peak traffic (50M)
Chaos engineering at high level
Gaurav is a Cloud Architect at Hotstar, India’s leading OTT provider that holds a global concurrency record for live streaming to 25Mn+ viewers. At Hotstar, he loves building cost-efficient infrastructure that can scale to millions in minutes. He is also passionate about chaos engineering and cloud security.