Introduction to kontinue

kontinue is a Kubernetes-native durable execution framework that lets you build resilient, long-running workflows. State is automatically persisted in Kubernetes resources so execution can tolerate crashes, node failures, and other infrastructure issues.

What is Durable Execution?

Durable execution ensures that your code runs to completion, even in the face of failures. Every step of your workflow is automatically persisted, allowing execution to resume exactly where it left off after any interruption.

This is essential for building reliable systems such as:

  • Multi-stage infrastructure provisioning
  • Long-running system processes
  • Complex deployment pipelines

Why kontinue?

Kubernetes Native

Unlike other workflow engines that require external databases, message queues, or services, kontinue is built from the ground up for Kubernetes. It uses etcd for state storage and Kubernetes primitives for coordination, resulting in:

  • Zero external dependencies
  • Seamless integration with existing Kubernetes infrastructure
  • Familiar operational model
  • Built-in high availability and fault tolerance

Developer Experience

Write workflows as regular Go code. No YAML configurations, no domain-specific languages, no code generation. You get full IDE support, type safety, and debugging capabilities.

func DeployWorkflow(ktx *kontinue.ExecutionContext, args *DeployArgs) (*DeployResult, error) {
    // Each step is automatically persisted
    kontinue.Execute[PrepareResult](ktx, PrepareInfraFn, &PrepareArgs{Cluster: args.Cluster}, &kontinue.ExecuteOptions{})
    kontinue.Execute[TestResult](ktx, TestInfraFn, &TestArgs{Cluster: args.Cluster}, &kontinue.ExecuteOptions{})
    kontinue.Execute[DeployResult](ktx, DeployFn, &DeployArgs{Cluster: args.Cluster, Version: args.Version}, &kontinue.ExecuteOptions{})

    return &DeployResult{}, nil
}

Production Ready

kontinue is designed for production workloads with:

  • Minimal overhead per workflow step
  • Automatic retries with configurable policies
  • Configurable timeouts
  • Comprehensive observability and tooling

Next Steps

Ready to get started? Check out our Quick Start guide to build your first durable workflow in minutes.