NUMAFLOW

Learn how to build and run real time streaming pipelines on Kubernetes with Numaflow, from first deployment to long term production operations.

Modern systems cannot wait for nightly batches. Teams need continuous analytics, fast reactions to events, and streaming workloads that scale with demand. Doing this on Kubernetes adds its own challenges, you must handle buffering, state, autoscaling, upgrades, and failure recovery without turning every pipeline into a custom distributed system.

This book walks you through Numaflow and the wider Numaproj ecosystem with a practical, operator friendly approach. You will see how Numaflow expresses pipelines as Kubernetes CRDs, how its control plane and data plane work, and how to design streams that stay correct under late data, bursts, and real failure modes.

• understand streaming fundamentals on kubernetes, including back pressure and delivery guarantees
• navigate numaflow architecture, control plane controllers, data plane containers, and isb services
• define pipelines with vertices and edges, including joins and complex topologies
• configure interstep buffering with nats jetstream for throughput or low latency needs
• work with event time, watermarks, windows, and state recovery for reduce operations
• write udf stages in go, java, python, and rust using the gRPC based process model
• use side inputs for dynamic rules and reference data with safe versioning practices
• choose between full pipelines, monovertex jobs, and servingpipeline for online inference
• deploy numaflow in clusters, size resources, and manage environments with gitops and argo cd
• operate in production with hpa and keda scaling, observability stacks, and debugging strategies
• integrate kafka, nats jetstream, and cloud pubsub sources and sinks for end to end use cases
• compare numaflow to flink, kafka streams, beam, and managed runtimes, then plan migrations and costs

This is a code heavy guide with working YAML and multi language UDF examples that you can adapt into real pipelines right away.

Grab your copy today and start shipping reliable streaming systems on Kubernetes.