Multi-Cluster Observability with Service Mesh - Challenges and Solutions

Explore multi-cluster observability with service mesh in this 35-minute conference talk by Ryota Sawada from UPSIDER, Inc. Delve into the complexities of observability in multi-cluster environments and examine the role of service mesh solutions. Learn from Sawada's experience running Istio in production since its v1.1 release, understanding both its benefits and challenges. Discover the limitations of service mesh as a silver bullet for complex observability requirements. Gain insights into cross-cluster tracing and metrics difficulties. Review Prometheus basics and understand Istio's default capabilities to identify gaps in observability. Address challenges in alert handling, high cardinality, and remote read/write operations. Conclude with a demonstration of a multi-cluster observability setup using Istio, Prometheus Operator, and Thanos.