Glossary · System coordination, integration and orchestration
Workflow orchestration
German: Workflow-Orchestrierung
In software and data engineering, workflow orchestration is the automated coordination of the tasks of a workflow, such as data jobs, service calls, machine steps or human approvals, by a central orchestrator that manages dependencies, scheduling, retries and monitoring.
- System integration
In one sentence
Workflow orchestration centrally coordinates workflow tasks, managing dependencies, scheduling, retries and monitoring.
Example
A workflow orchestrator runs the nightly pipeline that extracts machine data, cleans and aligns it, retrains a quality prediction model and publishes a report.
How it applies
- Engineering: Workflows are typically modeled as directed graphs of tasks with dependencies. Tasks should be idempotent, so retries after failures don't duplicate effects.
- Operation: The orchestrator provides visibility into what ran, what failed and why. Alerting on failed or delayed workflows is essential.
- AI context: Orchestrating AI steps, such as model calls, tool use and human review, is increasingly common; see Agentic workflow. Human approval steps remain important where outputs affect products or safety.
- Documentation: Documentation pipelines (building, translating and publishing content) are workflows too. Document each workflow's inputs, outputs, schedule, owner and failure handling.
Workflow orchestration vs. workflow engine
The Workflow engine is the software that executes workflows. Workflow orchestration is the activity or architectural approach of coordinating tasks centrally. A workflow engine is one tool for workflow orchestration.