How tmforum-oda/oda-canvas works
• The ODA Canvas Core – this is the central orchestration layer that manages overall lifecycle, initiates deployments and integrates with the various operators.
• Operator Subsystems – these are independent, loosely coupled services. They include:
– The Component Management operator (handles lifecycle and decomposition of a component into its sub-resources).
– The API Management operators (with separate implementations for different API Gateways or Service Meshes such as Istio, Kong, Apache APISIX, Apigee, Azure APIM, etc.).
– The Identity Config operator (for configuring identity management services).
– The Secrets Management operator (optional but present for secret configuration).
– The Dependent APIs operator (to assist with dependency discovery and integration).
• Canvas Portal – the user interface and back-end services for managing the canvas (visible in subdirectories like canvas-portal, which also include a web front-end, REST controllers, service layers, and helm configurations).
• CI/CD and Deployment Pipelines – deducible from extensive GitHub workflows, automation generators, Helm chart repositories (charts/*), and Dockerfiles. This illustrates building, releasing, and deploying various operators and base images.
• Supporting Tools & Test Frameworks – the feature definition-and-test-kit, automation scripts, and logging/observability configurations (e.g., Observability-design.md) round out the picture.
How these components interact
• The ODA Canvas Core communicates with each operator via API calls or messaging events (indicated by “Event-based-integration-design.md” and orchestrated workflows).
• The Canvas Portal serves as the UI/Service façade, interfacing with the back-end and indirectly invoking operators.
• The CI/CD pipeline (with GitHub actions workflows and automation generators) is responsible for building images, running tests, and deploying Helm charts.
• Modular integration: each operator’s deployment and lifecycle is controlled via chart installations (notice the extensive charts directory for various operators).
Architectural patterns and technologies
• Microservices/Multi-tenant design, where each operator is an independent service.
• Use of containers/Docker images as shown in numerous Dockerfiles and baseimages.
• Helm charts for Kubernetes management & deployment.
• Event-driven and RESTful APIs enforcing decoupling.
• CI/CD automation pipelines using GitHub Actions.
• The use of Jinja2 templating for workflow generation and scripting within automation components.
Final suggestions
• Clearly label every component with both its function and any pertinent technology (e.g., “Helm Chart Deployment for API Operator (Istio)”).