Cadence is a distributed orchestration engine that executes asynchronous long-running business logic in a scalable and resilient way.
Cadence addresses the challenge of reliably executing complex workflows across distributed systems where individual steps may fail, be retried, or span extended periods. The engine persists workflow state to a database (supporting Cassandra, MySQL, or PostgreSQL) and coordinates execution through multiple backend services. Workers running your workflow code communicate with the Cadence backend to receive tasks, execute business logic, and report results. The platform handles retries, timeouts, and state management automatically, allowing developers to write workflows as straightforward code rather than managing distributed coordination manually.
Teams should adopt Cadence when building systems that require durable, long-running processes with complex retry logic and failure handling. It suits microservice architectures where workflows orchestrate calls across multiple services, financial transaction processing, data pipelines, and any scenario where workflow state must survive service restarts. The project provides official SDKs for Go and Java, with community-maintained Python and Ruby libraries available. A web UI at localhost:8088 displays workflow histories and execution traces. Kubernetes deployment is supported through an official Helm chart, with guided installation available via KubeStellar Console. The CLI tool enables operational management of workflows, task lists, domains, and clusters.
The project maintains active continuous integration with automated testing. Code coverage is tracked and publicly reported. The team responds to community contributions through discussions and maintains multiple official client libraries alongside the core engine. Documentation includes sample implementations for both Go and Java, and the project accepts community-developed language bindings and framework layers built on top of the core platform.
