core-sdk

To build a Kailash workflow, use WorkflowBuilder to add nodes and connect outputs to inputs, then call build before execution. This ensures correct node configuration and valid data passing for reliable workflow execution.

Choose LocalRuntime for Kailash workflow execution in scripts and CLI work, and AsyncLocalRuntime for Docker and FastAPI deployments. Selecting the correct runtime ensures proper async execution and production-safe operation in your target environment.

Parameter passing in Kailash ETL pipelines allows you to pass runtime overrides into nodes during execution. This standardizes data flow and conditional routing, enabling flexible pipeline configurations without modifying the core node patterns.

Yes, you can integrate MCP and tracing instrumentation into Kailash workflows. The Skill provides canonical patterns for adding metrics, tracing, and observability instrumentation to ensure production-safe execution and effective debugging.

Handling cyclic flows and error handling in Kailash workflows requires applying specific node patterns to avoid execution failures. While the SDK supports these advanced patterns, incorrect configuration can break production-safe execution and disrupt async execution.