Temporal Workflow Engine

This open-source microservices orchestration platform enables developers to build scalable and reliable applications by managing distributed business logic as code, without the complexity of state machines, retry logic, and error handling. Originally developed by Maxim Fateev and Samar Abbas at Uber (as Cadence) in 2017, it was spun off as an independent company in 2019 and has over 12,000 stars as of 2026, with adoption by Snap, Stripe, Netflix, Datadog, and Coinbase. Key features include: workflow as code (defining business processes as regular code in Go, Java, TypeScript, Python, or PHP, with deterministic execution guaranteed by the runtime), activities (non-deterministic operations such as API calls, database queries, and file operations executed by worker processes with automatic retry, timeout, and compensation), durable execution (workflow state is automatically persisted and can survive process crashes, machine failures, and deployments, with replay-based recovery), timers and scheduling (long-running timers and cron-like scheduled workflows spanning days or weeks), signals and queries (external communication with running workflows via signals for state changes and queries for state inspection), child workflows (hierarchical composition with isolated execution and failure domains), versioning and patching (modifying workflow code without breaking running instances via version markers and patching APIs), search attributes (indexing workflows with custom attributes for filtering and listing), retry policies (configurable strategies with exponential backoff, maximum attempts, and per-activity overrides), timeout configuration (start-to-close, schedule-to-start, and heartbeat timeouts), worker pools (scaling workers horizontally across machines with task queue-based distribution), observability (metrics, tracing, and web UI for monitoring and debugging), and multi-language SDKs (Go, Java, TypeScript, Python, PHP, and .NET with cross-language interoperability).