Kestra

What Kestra does

Kestra is an orchestration platform: it starts, sequences and monitors the steps of a workflow, retries what fails and keeps a record of every run. Each flow is a YAML file listing its tasks and triggers. You can write that YAML in the built-in code editor, assemble the same flow in a no-code editor with a form for each task, or ask the AI Copilot to draft it. All three edit one definition, so the flow stays readable as code and can be kept in Git.

Triggers are where Kestra overlaps with general automation tools. A flow can run on a schedule, when a webhook is called, when a file lands in storage, or in real time when a message arrives on Kafka, MQTT, AMQP, NATS, SQS, Pub/Sub or another bus. Tasks come from a plugin catalog that covers databases, cloud storage, data tools such as dbt, notifications and generic HTTP calls, and script tasks run your own Python, Node.js or shell code. Backfills replay a schedule over a past period, which is useful for data pipelines.

Open source and Enterprise editions

The repository is Apache-2.0, and its Open-Source Edition has no limits on flows or executions. Kestra also sells a closed Enterprise Edition, enabled with a license key, and the differences matter for teams:

• Access control. The open source edition offers basic authentication only, which the documentation describes as suited to one person or a small team with shared credentials. SSO, LDAP, SCIM, role-based access control and audit logs are Enterprise features.
• Secrets. In the open source edition, the secret() function reads base64-encoded environment variables. Built-in secret storage and external managers such as HashiCorp Vault are Enterprise features.
• Scale. The open source edition runs on one server by default with a PostgreSQL or MySQL queue. Kafka, Redis or AMQP queues, high availability, worker groups and multi-tenancy are Enterprise features.

The pricing page says most plugins are available in both editions.

Self-hosting Kestra

The quickest start is a single Docker container. For a lasting install, the documentation provides a Docker Compose file with PostgreSQL. Kubernetes, AWS and Google Cloud deployment guides are also available. Kestra is a Java service, so expect a heavier footprint than a small Node.js or Go tool. The official Compose file mounts the Docker socket so that script tasks can run in their own containers, which gives the server broad control over the host. The same file notes that the management port is unauthenticated and leaves it unexposed by default.

Hosted version

Kestra Cloud is a fully managed service run by the company. At the time of review, access is by request from the pricing page.

Limitations

Kestra is built for data, platform and infrastructure engineers. It does not offer Zapier-style recipes where a non-developer picks two apps and maps fields. Most real flows involve YAML and some scripting. Teams that share one instance will quickly need SSO, roles or proper secret storage, and those are only in the paid edition.

Who it suits

Kestra suits teams that orchestrate data pipelines, scheduled jobs, infrastructure tasks and AI workflows and want them declared as code under an Apache-2.0 license. For code-first scripts with generated UIs, compare it with Windmill. For app-to-app business automation, start with Activepieces. The workflow automation guide explains the difference between these kinds of tool.