Skip to main content

Serverless Workflow

Serverless Workflow is a CNCF-hosted specification for defining serverless application workflows using a declarative, JSON- and YAML-based domain-specific language for orchestrating event-driven services.

  • JSON and YAML-based workflow definition language for serverless and event-driven applications (workflow orchestration)
  • Specification for states, control flow, and data passing in long-running service workflows (application integration)
  • Built-in constructs for event, function, and callback interactions across services (event-driven architecture)
  • Portable workflow model decoupled from any single runtime or cloud provider (multi-platform portability)
  • Extensible specification with pluggable function, event, and data mapping definitions (extensibility framework)

More About Serverless Workflow

Serverless Workflow is a Cloud Native Computing Foundation (CNCF) specification that defines a standard, declarative way to model serverless and event-driven business processes as workflows. It focuses on describing how distributed services, functions, and events interact over time, independent of any particular cloud provider, runtime engine, or programming language. The specification targets use cases where multiple event sources, functions-as-a-service, APIs, and long-running tasks must be coordinated in a repeatable and machine-readable manner (workflow orchestration).

At its core, Serverless Workflow provides a JSON- and YAML-based domain-specific language for describing workflows as a series of states and transitions (application integration). States can represent various control-flow patterns, including event consumption, function invocation, parallel execution, data-based decisions, and sub-workflow calls. Each state is formally defined in the specification, including its input and output data, transition rules, error handling behavior, and timeouts. This state model enables runtimes to execute workflows consistently across environments that support the specification.

The specification includes constructs for event-based interactions, function references, and data transformations (event-driven architecture). Workflows can subscribe to events, wait for them, or emit them as part of the process definition. Functions can be referenced by name and type, with details such as operation, endpoint, or resource information provided outside the core control-flow logic. Data mapping definitions enable workflows to select, transform, and merge data between states using declarative expressions. These capabilities align Serverless Workflow with categories such as cloud-native integration, serverless orchestration, and business process automation.

Serverless Workflow is designed as a portable model that separates the workflow definition from execution environments (multi-platform portability). The specification describes the language, syntax, and semantics but does not prescribe a specific engine implementation. This allows vendors, open-source projects, and internal platform teams to build runtimes that understand Serverless Workflow definitions while integrating with their own event brokers, function platforms, and service meshes. Enterprises can define workflows once and execute them on different compliant runtimes with minimal changes to the workflow description.

The project’s extensibility model allows function, event, and data mapping definitions to be customized and extended (extensibility framework). Workflows can reference external resources for function and event catalogs, security configuration, and operation metadata, which enables integration with existing Application Programming Interface (API) gateways, function registries, and event schemas. This design supports usage in complex enterprise environments that span hybrid and multi-cloud deployments, where standardization of workflow definitions can reduce coupling between applications and underlying infrastructure services.

From a directory and taxonomy perspective, Serverless Workflow fits into categories such as workflow definition language, serverless orchestration, event-driven integration, and cloud-native application coordination. It is relevant to enterprise architects, platform engineering teams, and application integration specialists who need a standardized way to describe long-running, event-driven workflows across heterogeneous platforms while retaining flexibility in runtime and infrastructure choices.