finds.dev← search

// the find

airflow-helm/charts

★ 708 · Shell · Apache-2.0 · updated Oct 2024

The User-Community Airflow Helm Chart is the standard way to deploy Apache Airflow on Kubernetes with Helm. Originally created in 2017, it has since helped thousands of companies create production-ready deployments of Airflow on Kubernetes.

A community-maintained Helm chart for deploying Apache Airflow on Kubernetes, predating the official Apache chart by several years. It targets teams that want a batteries-included Airflow deployment without writing their own Kubernetes manifests from scratch.

- Unusually thorough documentation: per-topic FAQ files covering pgbouncer, LDAP/OAuth, autoscaling, log persistence, and more — most charts dump everything into one README and call it done.

- PgBouncer is built in and ships its own maintained Docker image, which is the right call for Airflow's connection-heavy workload — the official chart makes you figure this out yourself.

- Sync deployments for connections, variables, pools, and users are declarative and GitOps-friendly, keeping Airflow config in values.yaml rather than requiring manual UI setup post-deploy.

- Supports CeleryExecutor, KubernetesExecutor, and CeleryKubernetesExecutor with separate sample values files, so you can see what a realistic production config looks like for each.

- Last pushed October 2024 and the bundled PostgreSQL image pins Bitnami PostgreSQL 11, which has been EOL since November 2023 — that embedded database is not safe to run in production.

- The official Apache Airflow Helm chart (apache/airflow) has largely caught up and is now the default recommendation from the Airflow project; this community chart risks becoming the legacy option as the official one matures.

- No Airflow 3.x support is visible in the directory tree or changelog, and given the pace of activity, it may never arrive — teams planning an Airflow 3 migration should not bet on this chart.

- The sync mechanism runs as separate Deployments and Jobs for each resource type (connections, variables, pools, users), which means a lot of moving parts to debug when something in that sync pipeline silently fails.

View on GitHub → Homepage ↗

// want more like this?

We dig through GitHub every week and send a few repos picked for what you actually care about — each with an honest take like this one.

Get finds in your inbox → Search again →