AWS App Mesh
By Amazon Web Services
AWS App Mesh is a managed service mesh that gives applications running on AWS consistent visibility and network traffic controls across microservices, regardless of the underlying compute platform.
Definition
AWS App Mesh is a managed service mesh that gives applications running on AWS consistent visibility and network traffic controls across microservices, regardless of the underlying compute platform.
Overview
App Mesh standardizes how services communicate by deploying an Envoy proxy — the same sidecar proxy used by Istio — alongside each service instance, whether it runs on Amazon ECS, EKS, Fargate, or plain EC2. Application code sends and receives traffic exactly as before, but every request now flows through the proxy, which App Mesh centrally configures with routing rules, retries, timeouts, and traffic-shifting policies defined through the AWS API, CLI, or CloudFormation. Because the mesh operates at the network layer independent of the compute platform, App Mesh lets teams standardize service-to-service communication across a heterogeneous fleet — for example, gradually migrating services from ECS to EKS without changing how they discover or call one another. It captures consistent metrics, logs, and traces for every hop in the mesh, which App Mesh exports to CloudWatch, X-Ray, and third-party observability tools, giving operators a single, protocol-aware view of service health across the entire application. App Mesh's core value is decoupling application logic from networking concerns: retries, circuit breaking, canary deployments, and traffic splitting are all configured declaratively in the mesh rather than hardcoded into each service. This puts it in the same category as Istio, Linkerd, and Consul — all implementations of the service mesh pattern — but App Mesh differentiates itself through deep native integration with AWS compute, networking, and observability services.
Key Features
- Envoy-based sidecar proxies for consistent service-to-service traffic control
- Works uniformly across ECS, EKS, Fargate, and EC2 workloads
- Declarative routing rules including weighted traffic shifting for canaries
- Automatic retries, timeouts, and circuit breaking configured at the mesh layer
- Native metrics, logging, and tracing integration with CloudWatch and X-Ray
- Virtual services, nodes, and routers as reusable abstractions for mesh topology
- TLS encryption between services managed through AWS Certificate Manager
- Gradual migration support across heterogeneous compute platforms