How orkes-io/orkes-conductor-client works
Project Type and Purpose
• This is an open-source Java SDK (client library) for the Orkes Conductor orchestration server.
• It is not a standalone application but a reusable client that:
• Exposes methods to configure and invoke Conductor workflows and tasks
• Provides pluggable HTTP and gRPC transports
• Integrates with secret managers (AWS, Azure) and external services (Kafka, AI integrations)
• Offers Spring Boot auto-configuration
File Structure
• Top-level buckets:
• src/ — main and test Java code
• examples/ — sample usage (helloworld, dynamic workflow, task workers)
• docs/ — markdown guides (AUTH, METADATA, worker and workflow patterns, Kafka queue)
• build.gradle, gradle/ wrapper/ — build and CI scripts
• .github/workflows — CI for tests, OSS publishing, release
• src/main/java split into two roots:
1. com.netflix.conductor.client/... — legacy Netflix Conductor client code
2. io.orkes.conductor.client/... — Orkes-branded client layers, including:
• ApiClient, OrkesClients façade
• HTTP layer under client/http (OpenAPI-generated API classes, interceptors, auth)
• gRPC layer under client/grpc (ChannelManager, pooled pollers, stream handlers)
• Automator (TaskRunner) and WorkerFn abstractions
• Model classes grouping API request/response DTOs
• SecretsManager abstractions for AWS/Azure
• Spring Boot auto-configuration
• Configuration & metadata:
• src/main/resources/service_config.json — default API endpoints
• Optional override via ApiClient constructor or Spring beans
README Insights
• Purpose: provide instructions to add SDK via Gradle/Maven, write “Hello World” workflows and workers.
• Demonstrates dynamic workflow creation, worker methods annotated with @WorkerTask, Main application launching WorkflowExecutor.
• Mentions Conductor server (Docker), external UI, and credentials (API key/secret).
Key Points to Capture
• Separation between transport layers (HTTP vs. gRPC) and common interfaces
• The polling-and-execution loop (TaskRunner) as the heart of the worker model
• How Workflow definitions map to client calls (WorkflowExecutor → Conductor REST API)
• Pluggable secret and integration clients, demonstrating extensibility
• Sample usage path: Main → ApiClient → WorkflowExecutor → worker methods → Conductor server