SYSTEM Cited by 2 sources
AWS Cloud Map¶
AWS Cloud Map is AWS's service-discovery registry: customers create a namespace (the logical boundary), register service instances under it, and clients resolve instance endpoints at query time. Used under the hood by App Mesh (as one of App Mesh's abstractions) and by Amazon ECS Service Connect (where the Cloud Map namespace is the mesh boundary, not just the service-discovery substrate).
Stub page — minimal viable for sources/2025-01-18-aws-app-mesh-discontinuation-service-connect-migration.
Role in App Mesh vs Service Connect¶
- App Mesh: one of four abstractions. Cloud Map namespace holds service-discovery records; the App Mesh Mesh resource is a separate logical boundary attached to the same namespace.
- Service Connect: single abstraction. The Cloud Map namespace is the mesh boundary — services attach directly, no separate Mesh resource.
This consolidation is a recurring theme in the App Mesh → Service Connect transition: fewer abstractions, same underlying primitives.
Constraint: no cross-account namespace sharing¶
Unlike App Mesh (which supports cross-account Mesh sharing via AWS RAM), Cloud Map namespaces cannot be shared across AWS accounts. In the Service Connect model this constraint propagates up: all tiers of an application must live in a single AWS Account.
Related¶
- systems/aws-app-mesh — historical primary customer
- systems/aws-ecs-service-connect — current primary customer
Seen in¶
- sources/2025-01-18-aws-app-mesh-discontinuation-service-connect-migration — service-discovery substrate in both App Mesh and Service Connect; the single-account constraint shapes the multi-tier-app story.
- sources/2026-04-08-aws-build-a-multi-tenant-configuration-system-with-tagged-storage-patterns — Cloud Map is the fan-out target resolver for event-driven config refresh: the invalidator Lambda queries Cloud Map at refresh time for healthy Config Service instances, then makes one gRPC refresh RPC per instance. Querying at refresh time (not caching the fleet list) is what makes the pattern robust to autoscaling / deployments / draining instances.