Health Check & Service Discovery  ·  View 01 of 21  ·  Context and scope

System Context

Three registration sources, one traffic path, and a list of things this plane is not permitted to need.

Editable source SVG draw.io All views
The traffic path it serves Internal callers 450 services Envoy sidecar fleet 60,000 subscriptions People Service owner SRE on call Security engineer Registration sources EKS control planes 12 clusters ECS + EC2 Auto Scaling VM and task fleets Service catalogue ownership, tiers Third-party endpoints no readiness signal Health & Discovery Plane Who exists · who gets traffic Dependencies and consumers Workload identity IAM roles for SA Route 53 + Cloud Map Observability platform declares health contract drains, evacuates audits denials pod + endpoint state instance state service of record declared, unprobed resolve by name versioned views authn DNS surface signals Health & Discovery Plane — System Context Application we own Person or role External / third party Security / platform synchronous Out of scope: load-balancing algorithms beyond the endpoint set and weights, mesh policy and mTLS, the public API edge. v 1.0 · owner Reliability Architecture · date 2026-10

Decisions

  • The plane is a dependency of nothing on the request path: callers route from a local cache, not from a call to this system.
  • Third-party endpoints are a registration source like any other, declared by their owner and never probed — capability is per service, not per platform.
  • The service catalogue, not this registry, is the system of record for what a service is and who owns it.

Assumptions

  • 450 internal services, 60,000 instances, 3 active regions, 12 M internal requests/second at peak.
  • 12 EKS clusters plus ECS and EC2 Auto Scaling fleets; a long tail of workloads the orchestrator does not own.

Deliberately out of scope

  • Load-balancing algorithms beyond the endpoint set and weights handed to them.
  • Mesh policy and mTLS planes, the public API edge, and the observability platform that stores the telemetry.
All views High-Level Architecture  ▶
Solution Architecture v1.0 · Amazon Web Services with an open-source Envoy/xDS data plane · Reliability Architecture · 2026-10 · 21 views · 16 architecture decision records01-system-context