Managed Airflow for dbt Orchestration

Yes. Datacoves’ Airflow supports the full Airflow operator library, including HTTP and Python operators, as well as custom decorators. Your DAGs can trigger external APIs, run Python scripts, kick off ML workflows, and integrate with tools such as Spark, dbt, and Airbyte within a single directed acyclic graph.

No, My Airflow provides a browser-based personal Airflow environment per developer with a single click.

Airflow’s built-in retry logic handles transient failures automatically. Datacoves layers in proactive alerting via email, Slack, or Teams so your team is notified the moment something fails, not when a stakeholder reports a stale dashboard. For critical pipelines, SLAs can be configured to escalate if a run does not complete within a defined window.

Both dbt retry and the result:error+ selector read run_results.json from the previous run, so the artifact has to survive the failure for either to work. Datacoves’ dbt-api stores it: a task declared with dbt_api_enabled=True and download_run_results=True pulls the previous run_results.json when it starts, then runs dbt build -s result:error+ --state logs if it finds one and a full build if it doesn’t. The retry rebuilds only the models that failed and everything downstream of them, in one task, with no per-model DAG to maintain. dbt-api is on by default; see retrying dbt tasks in Airflow for a full DAG.

Self-managed Airflow requires your team to provision infrastructure, manage Docker images, handle Kubernetes, configure secrets, and coordinate upgrades, all before you write a single DAG. Datacoves delivers a production-ready Airflow environment with managed upgrades, built-in secrets management, SSO, and per-user dev environments. The result: your data engineers spend their time building pipelines rather than operating Airflow.

My Airflow gives each developer a personal, isolated Airflow instance accessible from the browser with a single click. Developers can test DAGs against real connections and variables without affecting the shared production environment. This removes one of the most common friction points in Airflow development: waiting for a DevOps team to provision a local or staging environment before you can test your logic.

Prefect and Dagster, now one company after Prefect acquired Dagster Labs in July 2026, are general-purpose orchestrators that can run dbt, but they require additional configuration, custom adapters, and separate infrastructure decisions. Datacoves is specifically built around the dbt development workflow, it ships custom dbt Airflow decorators, virtual environments pre-configured for dbt, and a development environment (VS Code + My Airflow) that mirrors production. Teams get an end-to-end platform rather than assembling one from parts.

Airflow sends failure alerts to email, Slack or Microsoft Teams out of the box. Grafana is also available for dashboards and threshold-based alerting. Because orchestration covers the whole pipeline, an alert can fire on an ingestion or downstream step as well as on a dbt run.