Airflow 2 lost open-source security support on April 22, 2026. Read about the migration

Airflow 2 is out of support. Migration details

Run airflow version. If it starts with a 2, we should talk.

Open-source Apache Airflow 2 stopped getting security patches on April 22, 2026. We move small teams to Airflow 3 at a fixed price, on MWAA, Cloud Composer, Astro, or self-hosted. The assessment is a fixed fee, credited against the migration.

Where Airflow stands

Version Status Date
Airflow 2.x, open source End of life April 22, 2026
Airflow 2.x on Astronomer (Astro) Extended support Until April 2027
Airflow 3.x Supported Since April 2025

Sources: Apache Airflow version policy; Astronomer’s Airflow 2 end of life page.

It’s a migration, not a version bump.

  • Core operators moved packages, so imports change across your DAGs.
  • Scheduling and dates changed. Jobs can succeed and still land data on the wrong date.
  • The metadata database needs a backup, a migration, and a rehearsal.

From assessment to cutover

  1. Assessment

    1 to 2 weeks, fixed fee

    A DAG inventory and a migration plan, with cutover and rollback. Ends with a fixed quote. The fee is credited against the migration if you go ahead.

  2. Migration

    Fixed price

    We convert DAGs and providers, prove parity, and run a rehearsed cutover.

  3. Operations

    Monthly, optional

    We keep it running and current as part of pipeline operations.

Priced by platform size, not hours: DAGs, environments, and where it runs.

2026

Airflow 2 to 3 migration

A fintech running its data pipelines on Astronomer.

DAGs
120
Platform
Astronomer
Duration
2 months

What broke

Mostly parameter checks. Everything else needed routine changes.

How parity was proven

Both versions ran side by side for 2 weeks, with outputs compared in Great Expectations.

Read the full case note

Questions

Do you need access to production?

Read-only for the assessment. Write access to staging for the migration.

Can our team do part of the migration?

Yes. Your team converts some DAGs, and we review every change.

What happens to running DAGs during cutover?

After a parallel run, we pause in a quiet window, migrate the database, and resume on Airflow 3. Airflow 2 stays ready as a rollback.

Know the plan before you move.

A free 30 minute call tells you whether you need an assessment, and what it would cover.

You’ll talk to John García, a data engineer with over 15 years building production pipelines.

Rather write? Leave a note and we’ll reply by email.