[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f3ca8ivw3pwkyv":3},{"code":4,"message":5,"data":6,"timestamp":57},10000,"success",{"full_name":7,"position":8,"description":9,"language":10,"stars":11,"forks":12,"stars_gained":13,"age_days":14,"hn_threads":15,"one_liner":28,"summary_md":29,"topic":30,"github_topics":33,"license":53,"homepage":54,"updated_at":55,"history":56},"kestra-io\u002Fkestra",14,"Orchestrate everything, everywhere. The open-source control plane for workflows across data, infrastructure, business processes and AI agents. Declarative YAML, event-driven, any language, Git-versioned, scalable to millions of executions.","Java",29977,3471,1750,2618,[16,22],{"id":17,"title":18,"points":19,"comments":20,"posted_at":21},30790047,"Show HN: Kestra - Open-Source Airflow Alternative",142,69,"2022-03-24",{"id":23,"title":24,"points":25,"comments":26,"posted_at":27},30670034,"Kestra: Infinitely scalable open source orchestration and scheduling platform",7,2,"2022-03-14","A declarative, language-agnostic orchestration control plane for scheduled and event-driven data, infrastructure and AI workflows.","## The problem\n\nOnce a team has more than a handful of scheduled jobs, the orchestration layer becomes the hard part: cron entries scattered across servers, scripts that fail silently, no single place to see what ran, what retried and what is blocked downstream. Kestra addresses this by making every workflow a versioned YAML definition executed by a central control plane with triggers, retries, error handling and observable outputs. It is not just a scheduler upgraded; it aims to be the plane that sits above all automation, whatever the underlying runtime.\n\n## What's different\n\nThe usual choices in this space are Airflow and its Python-DAG model, or glued-together cron plus CI. Kestra's distinctive bet is that the flow definition is declarative YAML with no host-language code required, so the same flow can call Python, shell, SQL, containers or cloud APIs without the orchestrator imposing a programming language on you. The second differentiator is that the UI, a no-code editor, the AI Copilot and Git all edit the same YAML revision, so visual editing does not fork from infrastructure-as-code. Third, 2.0 moves workers to a gRPC worker-controller split from the control plane, which matters for running workers in another region or a restricted network.\n\n## Who it's for\n\nData platform and platform-engineering teams that already run Airflow, Dagster or a pile of cron jobs and want a single control plane covering data pipelines, infra automation and, increasingly, LLM workflows. It fits organisations that care about Git-versioned workflows and namespaces separating teams. Teams that only need a single daily script should not bother; a cron job is cheaper. Teams deeply invested in Python-native DAG authoring will find Kestra's YAML-first model more indirect than Airflow's code-first one, and pure CI\u002FCD pipelines are usually better served by their existing CI system.\n\n## Cost of adoption\n\nKestra is a Java server with a database and separate worker processes, so you are operating a real service, not a library: expect container or VM deployment, persistent storage and PostgreSQL-class database work. The plugin ecosystem is broad but versioned separately, and the 2.0 release carries breaking changes (for example ForEach replaced by Loop, worker communication moved to gRPC), so upgrades need the migration guide. The README is rich on concepts and light on resource sizing and failure modes at scale, so a proof of concept before production is warranted. Apache-2.0 keeps licensing unproblematic.\n\n## Verdict\n\nThis is an established project, not a spike: created in 2019, roughly 30k stars, and a monthly Java board appearance with about 1,750 stars gained in the tracked month shows interest still rising rather than fading. The 2.0 architecture work and the AI-native additions suggest active, opinionated maintenance. The trade-off is that you are adopting an operational platform with its own abstractions, so the decision hinges on whether you genuinely have cross-domain workflows to unify. If you do, Kestra is a credible alternative to Airflow that avoids Python-DAG lock-in; if you do not, it is infrastructure you did not ask for.",{"slug":31,"name":32},"devops","DevOps & Self-hosting",[34,35,36,37,38,39,31,40,41,42,43,44,45,46,47,48,49,50,51,52],"ai-agents","automation","control-plane","data-engineering","data-orchestration","data-orchestrator","etl","hacktoberfest","high-availability","infra-automation","infra-ops","infrastructure-as-code","java","low-code","orchestration","pipeline","pipeline-as-code","scheduler","workflow","Apache-2.0","https:\u002F\u002Fgo.kestra.io\u002Fhome","2026-10-24T17:36:26Z",[],1792871968]