The Pipeline Framework

Appearance

Orchestrator Runtime ​

The Pipeline Framework generates a single orchestrator runtime that coordinates the pipeline flow. It provisions inputs, invokes step clients, and handles the final outputs.

Canvas First
Configure the orchestrator flow in the Canvas at https://app.pipelineframework.org, then customize the generated runtime as needed.

What It Does ​

  1. Initiates pipeline execution
  2. Provides input data
  3. Coordinates step-to-step flow
  4. Handles final outputs

Generated Structure ​

text
orchestrator-svc/
├── pom.xml
├── src/main/java/
│   └── com/example/app/orchestrator/
│       └── OrchestratorApplication.java
└── src/main/resources/
    └── application.properties

The generated OrchestratorApplication is annotated with @PipelineOrchestrator so the annotation processor can generate the server endpoint (REST or gRPC, based on pipeline-config.yaml) at compile time.

Input Options ​

The generated runtime supports several ways to specify input:

  1. CLI argument: -i / --input
  2. Environment variable: PIPELINE_INPUT
  3. Quarkus config: quarkus.pipeline.input
  4. System property: -Dquarkus.pipeline.input=/path

The CLI argument takes precedence, followed by environment variable and config sources.

Customizing Input ​

Implement the getInputMulti() stub in the generated OrchestratorApplication to map inputs to domain types and feed the pipeline.