Real-time Kubernetes - How Entain Australia 10x'd Throughput with Linkerd

Discover how Entain Australia revolutionized their Kubernetes deployment in this insightful conference talk. Explore the journey of a large-scale system with 300 microservices and 3000 pods across multiple clusters and regions. Learn about the transition from connection-level to request-level load balancing, resulting in a tenfold increase in application throughput through the implementation of Linkerd. Gain valuable insights into the impact of Linkerd's load balancing on gRPC calls, the decision-making process for selecting a service mesh, and practical lessons from deploying a service mesh in a production environment. Delve into Entain's microservices journey, addressing growth challenges, service discovery evolution, and the unexpected benefits of their chosen solution. This presentation offers a comprehensive look at how a real-time Kubernetes system can be optimized for maximum performance and efficiency.

Intro
About Entain Australia
Steve Gray
Our Software Stack
Our Platform is Cloud Native
Some Key Metrics
It Wasn't Always This Way....
Our Microservices Journey
The Challenges of Growth
Service Discovery Evolution
GRPC Woes....
Rolling Out Code.....
C'est La Vie?
It Went Poorly
Days of Effort Lost
Solution Selection
Our Selection
The Deployment
How? What? How is it....
Unexpected Gains
Other Features We Got....
Our Journey Onward