¶ Fuse Pipeline Component Framework
The Pipeline Component Framework is a foundational part of the eTag Fuse platform. It provides a modular, governed runtime for building reusable components that operate within orchestrated pipelines. These components support integration, automation, and cross-domain execution in a consistent, traceable manner.
Components in Fuse are not standalone scripts — they are interoperable, policy-aware units of execution within the orchestration layer.
¶ 1. Overview and Core Purpose
-
Composable Execution:
Design reusable components that participate in pipelines across workflows and domains. -
Separation of Concerns:
External integrations (connectors) and internal logic (steps, adapters, utilities) are clearly separated. -
Governed Runtime:
Components execute under the control of identity, role, and runtime policies using scoped access. -
Scalable Orchestration:
Pipelines composed from components can scale to handle diverse and high-volume workloads. -
Traceable Processing:
Message context, input/output, and lifecycle transitions are observable and logged.
¶ 2. Key Features
-
Component Type System:
Components are categorized usingComponentTypeTypeflags to describe their purpose:- Connector — Interfaces with external systems, APIs, or services.
- Adapter — Performs data transformation or enrichment.
- Process — Executes a logical or functional step in the pipeline.
- Utility — Supports runtime or orchestration logic without modifying the message.
- Experimental — Optional flag indicating beta or in-development components.
-
Message-Based Execution:
Components operate on structured messages that carry a payload, metadata, and transient context. -
Metadata-Driven Discovery:
Fuse automatically discovers and registers components using attributes like[Component],[Method],[Property], and[MethodArgument]. -
Expression Support:
Dynamic expressions drive routing, control logic, validations, and conditions at runtime. -
Error Management:
Built-in support for retries, failovers, fallback logic, and compensation steps. -
UI Auto-Generation:
Component attributes define how inputs and settings are rendered in the Fuse interface.
¶ 3. Component Lifecycle
Components in Fuse follow a well-defined lifecycle:
Uninitialized > Initializing > Initialized > Starting > Started
> (Paused / Resuming / Stopping / Stopped / Disposed / Error)
You can override lifecycle methods such as:
OnInitialize()OnStart()OnPause()OnResume()OnStop()OnDispose()
These methods enable resource setup, teardown, and status transitions aligned to orchestration state.
¶ 4. Message & Variable Handling
- Payload — The primary data object processed (e.g., JSON, binary, stream, entity)
- Metadata — Routing and execution info like timestamps, retry counts, or step names
- Context — Transient step-specific values (e.g., calculated fields or temporary flags)
- Pipeline Variables — Persisted across components; accessible for conditionals and expressions
All components must declare a Message parameter as the first method argument to operate properly in the pipeline execution model.
¶ 5. Control Flow
Pipeline orchestration supports:
Sequential steps
Parallel branches
Conditional paths using expressions- Loops and iteration logic
Error handling and recovery flows
Human-in-the-loop approvals and gates
Control flow metadata can be enforced using MethodTypeType, MethodOptionType, and runtime variable bindings.
¶ 6. Scheduling & Commands
-
Scheduling Options:
- Time-based (cron, interval, time zones)
- Event-based (webhooks, messages, file triggers)
- Manual or programmatic invocation (API, UI)
-
Component Runtime Commands:
Start,Pause,Stop,Resume,Dispose- Trigger retries or reset component state
- Dynamically reconfigure component properties
Commands are often used in operational environments for fault recovery or controlled restarts.
¶ 7. Logging & Monitoring
-
Structured Logs:
Every component emits logs with trace IDs, execution time, payload state, and result. -
Log Levels:
Supports standard levels: Trace, Debug, Info, Warning, Error, Critical -
Monitoring:
Logs and telemetry feed into the Observability Domain and can be:- Viewed in real-time dashboards
- Analyzed for retry patterns and failures
- Routed to external log aggregators or SIEM tools
¶ 8. Supported Integration Patterns
Component metadata defines one or more IntegrationPattern flags that indicate supported orchestration use cases, including:
PipelineAPIApplicationETLStorageDataDeviceNotificationAILoggingSecurityDocumentsMarketsProvisioningMessagingCRMWebKnowledgeBaseDiscoverySynchronizationEnvironment
These patterns allow Fuse to surface components appropriately in UI builders and restrict them to valid orchestration contexts.
¶ 9. Development & Reuse
Components can be developed using the following frameworks:
- Method Framework — Define logic blocks and execution steps
- Entity Framework — Persist reusable, structured objects
- Property Framework — Define inputs, metadata fields, and typed configurations
- Artifact Package Framework — Bundle and deploy versioned components across environments
See How to Create Fuse Components for full implementation guidance, attribute references, examples, and lifecycle details.
¶ 10. Best Practices Summary
- Always annotate public methods, arguments, and properties
- Begin all method signatures with a
Messageparameter - Use partial classes for large or multi-functional components
- Provide full XML summaries and descriptions for every method
- Use
IntegrationPatternandComponentTypeTypeto clearly define orchestration role - Handle all payload types explicitly (e.g.,
byte[],MemoryStream,Entity) - Avoid hard-coded logic — use pipeline variables and expressions
Next: Method Framework — Register custom logic and expressions to use within pipeline components.