name: Service Fabric Vocabulary description: >- Domain vocabulary for Azure Service Fabric, the open-source distributed systems platform for building scalable and reliable microservices and containerized applications. version: "9.1" created: '2026-05-02' modified: '2026-05-02' terms: # Core Platform - term: Cluster definition: >- A network-connected set of virtual or physical machines (nodes) into which microservices are deployed and managed. Service Fabric clusters have a control plane (cluster manager) and can span availability zones for high availability. related: - Node - Cluster Manager - term: Node definition: >- A machine (physical or virtual) that is part of a Service Fabric cluster. Nodes are assigned a node type and can host many application replicas. Seed nodes participate in cluster quorum. related: - Cluster - Node Type - Seed Node - term: Node Type definition: >- A named group of nodes in a cluster sharing the same VM SKU, OS, and port configurations. A cluster can have multiple node types. The primary node type hosts system services. related: - Node - term: Seed Node definition: >- A special cluster node that participates in leader election and forms the quorum for cluster management decisions. Seed nodes must be up for the cluster to be operational. related: - Node - Quorum # Application Model - term: Application definition: >- A collection of constituent services that together form a composite application deployed to a Service Fabric cluster. Applications are instances of application types and are identified by a fabric:/ URI. related: - Application Type - Service - term: Application Type definition: >- A named and versioned collection of service type definitions. An application type is defined in an ApplicationManifest.xml and must be provisioned to the cluster before instances can be created. related: - Application - Service Type - term: Service definition: >- A deployable, versioned component of an application. Services implement business logic and are either stateless or stateful. Service Fabric manages service placement, replication, and health. related: - Service Type - Stateless Service - Stateful Service - term: Service Type definition: >- The typed definition of a service as declared in a ServiceManifest.xml. Defines the code, configuration, and data packages that constitute a service. related: - Service - Application Type - term: Stateless Service definition: >- A Service Fabric service where each instance is identical and no state is maintained across requests. Multiple instances can be deployed for load balancing. State is stored externally (database, cache). related: - Stateful Service - Instance - term: Stateful Service definition: >- A Service Fabric service that maintains durable state using Reliable Collections or Reliable Actors. State is replicated across multiple replicas for high availability. related: - Stateless Service - Reliable Collections - Primary Replica # Partition and Replication - term: Partition definition: >- A subset of a service's total state managed by one set of replicas. Partitioning enables horizontal scaling — different nodes manage different partitions. Partitioning schemes: Singleton, UniformInt64Range, Named. related: - Replica - Service - term: Replica definition: >- An instance of a service partition running on a node. For stateful services, one replica is the Primary and others are Active Secondaries. Replicas are managed by the Failover Manager. related: - Partition - Primary Replica - Secondary Replica - term: Primary Replica definition: >- The replica responsible for processing read and write requests for a stateful service partition. If the primary fails, a secondary is promoted automatically. related: - Replica - Secondary Replica # Programming Models - term: Reliable Collections definition: >- Distributed, replicated data structures (Reliable Dictionary, Reliable Queue) provided by the Service Fabric SDK. They behave like standard .NET collections but are replicated across replicas for durability without an external database. related: - Stateful Service - Reliable Services - term: Reliable Actors definition: >- An actor-model programming framework for Service Fabric. Each actor is an isolated, single-threaded object with its own state. Actors communicate via asynchronous messages and are automatically placed on cluster nodes. related: - Stateful Service - term: Reliable Services definition: >- The base programming model for building both stateless and stateful services in Service Fabric. Provides the RunAsync lifecycle, communication listeners, and Reliable Collections access. related: - Stateless Service - Stateful Service # Health Model - term: Health State definition: >- A Service Fabric entity's health status as computed by the Health Manager. Values: Ok (healthy), Warning (degraded but functional), Error (unhealthy), Unknown (health unknown), Invalid. values: - Ok - Warning - Error - Unknown - Invalid related: - Health Report - Health Policy - term: Health Report definition: >- A health evaluation submitted by a watchdog, service, or operator describing the health of a cluster entity (node, application, service, partition, replica). Reports include source, property, health state, description, and TTL. related: - Health State # Deployment and Upgrade - term: Rolling Upgrade definition: >- Service Fabric's default upgrade mode that updates services incrementally across upgrade domains. Monitors health at each step and can automatically roll back if health degradation is detected. related: - Upgrade Domain - Application - term: Upgrade Domain definition: >- A group of nodes that are upgraded together during a rolling upgrade. Service Fabric pauses and evaluates health between upgrade domains. Typically aligned with fault domains. related: - Rolling Upgrade - Fault Domain - term: Fault Domain definition: >- A set of nodes that share a common point of failure (same rack, power unit, or network switch). Service Fabric distributes replicas across fault domains for high availability. related: - Upgrade Domain # Communication - term: Naming Service definition: >- A Service Fabric system service that resolves service names to endpoint addresses. Applications use the Naming Service to discover service endpoints before making calls. related: - Service Fabric DNS Service - term: Service Fabric DNS Service definition: >- An optional cluster service that assigns DNS names to Service Fabric services. Enables services to be resolved by DNS name (e.g., myservice.myapp) rather than fabric:/ URIs. related: - Naming Service categories: - name: Platform terms: - Cluster - Node - Node Type - Seed Node - name: Application Model terms: - Application - Application Type - Service - Service Type - name: Service Types terms: - Stateless Service - Stateful Service - name: Replication terms: - Partition - Replica - Primary Replica - name: Programming Models terms: - Reliable Collections - Reliable Actors - Reliable Services - name: Health terms: - Health State - Health Report - name: Deployment terms: - Rolling Upgrade - Upgrade Domain - Fault Domain - name: Communication terms: - Naming Service - Service Fabric DNS Service