Mastering Chaos - Achieving Fault Tolerance with Observability-Driven Prioritized Load Shedding

Explore the challenges of microservices-based applications and learn about Aperture, an open-source tool for observability-driven prioritized load shedding, in this 40-minute conference talk from SREcon23 Asia/Pacific. Dive into the complexities of metastable failures like cascading failures and retry storms, and discover how Aperture addresses the shortcomings of current approaches. Examine Aperture's innovative architecture, including its control and data planes, and understand how it utilizes token buckets, weighted fair queuing, and concurrency limiting to effectively prioritize workloads. Gain insights from real-world implementations of Aperture in cloud products, showcasing its ability to protect multi-tenant databases from overloads through prioritized load shedding of GRPC and GraphQL traffic. Uncover a powerful solution for ensuring the reliability and resilience of microservices-based applications in this informative presentation by Harjot Gill and Hardik Shingala from FluxNinja, Inc.