Using Hystrix to Build Resilient Distributed Systems

Track: Microservices
Skill Level: Intermediate
Room: Room A301
Time Slot: Wed 2/17, 10:30 AM
Tags: java , hystrix , distributed systems , microservices , jvm

Hystrix is an open-source library developed by Netflix to help us build resilient distributed systems. It provides boundaries that prevent failures in one service from cascading to other services. It is also designed with the knowledge that failures in a distributed system are inevitable, and have to be handled as a first-class citizen through the use of fallbacks. Used properly, Hystrix will greatly improve the availability of a system that may encounter errors or increased latency in a distributed environment.

In my talk, I will describe the principles of Hystrix and what role it plays in operating the Netflix API. I will describe a variety of cases where proper Hystrix usage has protected our API, and lessons we learned about how improper usage can lead to customer-affecting outages.

Matt Jacobs

Matt is an engineer on the Edge Platform team at Netflix, responsible for the availability of the Netflix API. He is the maintainer of Hystrix, and applies it to a variety of systems at Netflix to help maintain a fault-tolerant system.