How controlm/automation-api-quickstart works
Project type and purpose
• This is not a monolithic application but a “samples repository” demonstrating how to use the Control-M Automation API.
• It provides end-to-end examples (JSON job-flow definitions, scripts in Python/Node/Powershell/SQL, Java JUnit tests, Dockerfiles, Kubernetes manifests).
• Two major categories at the top level:
– control-m – samples for the standard Control-M Automation API
– helix-control-m – samples for the Helix Control-M API flavor
File structure
• Top-level: LICENSE, README.md, two folders (control-m, helix-control-m).
• Under each category: numbered folders 101, 102, 103,… grouping samples by capability (first job, file transfer + DB, Hadoop/Spark, script jobs, code deployment, Docker, onboarding, JUnit, DEV→PROD, corrective flows, REST examples, CSV→JSON, Kubernetes agent, monitoring).
• Inside each sample folder:
– A README.md explaining the sample
– One or more definitions (.json for flows, descriptors, Kubernetes YAML, Dockerfiles)
– Scripts or code in Python, Shell, PowerShell, Node, Java, SQL
• No large server-side code: everything is client-side artifacts that invoke the Automation API to drive a Control-M environment.
Key architecture insights from README
• This is a quickstart/demo for the Control-M Automation API.
• Users need: a local Docker Workbench image (Control-M environment), the ctm-cli Automation API CLI, and connectivity to a Control-M server.
• Samples show how to deploy job definitions, run jobs, query status, build agents for Docker or Kubernetes.
Main components
• Client workstation (developer’s laptop)
– Sample code folders (control-m / helix-control-m)
– Automation API CLI (ctm-cli) running locally or in Docker
• Control-M server environment
– Control-M server REST API endpoint
– Scheduler and database
• Control-M agents
– Traditional agents (installed on hosts)
– Docker agents (built from Dockerfiles in samples)
– Kubernetes agents (statefulset, YAML manifests)
• External execution targets
– File servers (for file-transfer jobs)
– Databases (for SQL jobs)
– Hadoop/Spark cluster (for big-data jobs)
– Generic compute (shell/PowerShell/Node/Python)
• Monitoring / corrective flows
– Optional external monitoring listener (alertsListener.ts sample)
– Sample corrective job flows
Relationships and data flows
• Developer edits JSON or script → invokes ctm-cli (or uses sample code via REST libraries / JUnit) → sends HTTP calls (POST/PUT) to Control-M REST API → Control-M server persists job-flow definitions in its database → Scheduler triggers jobs at runtime → Agents pick up job steps → Jobs execute on external resources → Agents report status back → Control-M server updates state → CLI / sample code polls GET /status.
Important architectural patterns
• Separation of concerns
– Client side (samples, JSON, scripts) vs server side (Control-M components) vs execution side (agents + targets)
• API-driven deployment
– All artifacts are declarative (JSON, YAML) and deployed by REST calls
• Extensibility
– Demonstrated via multiple language bindings (CLI/shell, Python, Node, PowerShell, Java)
• Containerization
– Dockerfiles to package agents
– Kubernetes manifests for stateful agents
Checklist
• All three layers with clear separation
• Docker build contexts (samples that contain Dockerfile), Kubernetes manifests (stateful_ha.yaml)
• Technology stacks alongside components (Java JUnit, Node/Axios, PowerShell, Python, SQL)