Intelligent service discovery

graph TD Client["Client Application"] SD["Service Discovery (Foundation.Distributed.Discovery)"] SR["Service Registry (Distributed & Health-Aware)"] subgraph "Registered Service Instances" Service1["Service Instance 1 (NodeX, Healthy, Caps: [:read_replica])"] Service2["Service Instance 2 (NodeY, Degraded, Caps: [:read_replica])"] Service3["Service Instance 3 (NodeZ, Healthy, Caps: [:other_cap])"] end %% Client requests service Client -- "Request: Find Service (Type: :database, Capability: :read_replica)" --> SD %% Service Discovery queries Registry SD -- "Query(criteria)" --> SR %% Registry shows available (but not yet filtered) services SR --> Service1 SR --> Service2 SR --> Service3 %% Registry returns results (conceptually, SD will filter these) %% SR -- "Results[Service1, Service2, Service3]" --> SD %% Mermaid doesn't easily show data packets on return arrows like this without clutter, %% so the filtering logic is shown in the next step by SD. %% Service Discovery selects the best match SD -- "Filter & Select Best Match (Healthy & Capable)" --> Service1 %% Service Discovery returns the selected service to Client SD -- "Response: Connect to Service Instance 1 @ NodeX" --> Client %% Client connects to the chosen service Client -.-> Service1 %% Styling classDef client fill:#E6E6FA,stroke:#333,color:#000; classDef discovery fill:#ADD8E6,stroke:#333,stroke-width:2px,color:#000; classDef registry fill:#FFFACD,stroke:#333,color:#000; classDef service fill:#98FB98,stroke:#333,color:#000; classDef unhealthy_service fill:#FFB6C1,stroke:#333,color:#000; classDef nonmatch_service fill:#LIGHTGRAY,stroke:#333,color:#000; class Client client; class SD discovery; class SR registry; class Service1 service; class Service2 unhealthy_service; class Service3 nonmatch_service;