Service, (Un)interrupted: How We Made a Non-EC2 Component Highly Available

In the book "High Availability: Design, Techniques, and Processes," Floyd Piedad emphasizes the importance of system availability from the user's perspective. A highly available system delivers operational performance consistently over a given period of time. Three principles of reliability engineering help achieve this: removing single points of failure, reliable crossover to redundant resources, and early detection of failure points. The case study by BrowserStack demonstrates how these principles were applied to make a non-AWS component highly available. By adding redundancy, implementing health checks, and using Route 53 for configuration, the system achieved inter-and-intra-region high availability while also benefiting from load balancing on Tweaker machines.