One post tagged with "observability"

If you havent already worked with Langflow, stop right now, download it immediately, and marvel at your newfound AI workflow productivity! Building a Langflow workflow is intuitive. Drag components, connect edges, test in the UI. Langflow even has tools that help you get to production. These tools allow you to run the flow without the Langflow UI, monitoring the execution, etc. But, when you look to scale that flow for production use, your only choice is distributing entire flows across Langflow replicas. This is where Stepflow, a general purpose workflow orchestration and distributino platform steps in. With Stepflow deployed to Kubernetes, we can run any Langflow workflow with full observability and distribution of steps within a flow across workers.

To showcase this, we've created a real-world example using an off the shelf Langflow flow. This post walks through running the Langflow workflow on Kubernetes using Stepflow's example production architecture. To do this, we'll convert a sample document ingestion pipeline workflow currently in use in the OpenRAG project. With Stepflow deployed to a local Kind cluster, we'll run this workflow and trace the execution through the system using Stepflow's built in telemetry stack.