Cloud Native Theater | EnvoyCon: Inside Spotify’s Envoy Architecture: What We Le... Leonardo da Mata

This talk, presented by Leonardo da Mata from Spotify, delves into the challenges and solutions surrounding the deployment of Envoy at scale. The speaker discusses the transition from a heavy Java sidecar to using Envoy-native custom C++ filters, highlighting key decisions made during the migration to Kubernetes and standardization on gRPC. The session covers the implications of these changes on latency, reliability, and operational costs, providing an honest account of both successful strategies and lessons learned for managing request-handling logic in a cloud-native environment.