Dagster

Dagster is a modern data orchestrator centered on assets rather than tasks, enabling teams to build, observe, and deliver reliable pipelines. Its 'asset-centric' approach automatically tracks data lineage and quality, showing teams what broke, why, and what depends on it, far beyond simple job success/failure reporting. Dagster+ AI enriches the experience by leveraging operational context (assets, runs, lineage, freshness) to aid diagnosis and more confident action. The platform offers branch deployments for safely testing pipeline changes, a hybrid model (local compute, Dagster control plane), and native integrations with dbt, Snowflake, and Fivetran. Pricing ranges between Solo ($10/month plus $0.04/credit, 30d free) and Starter ($100/month plus $0.035/credit, 30d free), with a Pro plan on request for enterprise. One credit equals one asset materialization or task execution; serverless compute costs an additional $0.010/minute.

Dagster does not provide a French-language interface and remains exclusively English-language. The REST API and plugin ecosystem enable deep integration with existing data warehouses and ELT pipelines, particularly suited for modern data stacks (dbt, Snowflake, Fivetran). The credit-based billing model proves transparent and fair for moderate pipelines, but can become expensive for massive executions without optimization. The learning curve is moderate for data engineers comfortable with Python and DAG concepts, but requires understanding modern data engineering patterns. The major strength is lineage and quality visibility that exceeds competing orchestration tools.