managing-astro-local-env_skill

Overview

This skill manages a local Airflow environment using the Astro CLI. It provides commands and guidance to start, stop, restart, inspect logs, troubleshoot containers, and reset or upgrade the runtime. Use it to keep a reproducible local Airflow instance for development and testing. It complements project setup and DAG authoring/testing skills.

How this skill works

The skill maps common Astro CLI commands to practical maintenance tasks: starting and stopping containers, checking container status, viewing and following logs, and opening shells inside containers. It advises on restart triggers (changes to dependencies or Dockerfile), offers troubleshooting steps for common failures, and provides a safe reset sequence. It also outlines how to test and change Airflow versions via the runtime Dockerfile.

When to use it

• Start a local Airflow instance for development or manual testing
• Stop or restart services after dependency or Dockerfile changes
• Inspect logs or follow real-time output to debug scheduler/webserver issues
• Run Airflow CLI commands or open a shell inside containers
• Reset the environment when containers are unhealthy or corrupted

Best practices

• Restart Airflow after modifying requirements.txt, packages.txt, or the Dockerfile
• Use astro dev ps to confirm component status before debugging
• Follow logs (astro dev logs -f) while reproducing issues to capture errors
• Use astro dev kill then astro dev start for a clean reset when containers fail to recover
• Test Airflow version upgrades with astro dev upgrade-test before changing the Dockerfile

Example use cases

• Start local Airflow and confirm webserver is available at http://localhost:8080
• Follow scheduler logs to diagnose missed DAG runs: astro dev logs --scheduler -f
• Open a shell in the scheduler container to run airflow CLI commands: astro dev bash; astro dev run airflow dags list
• Reset a broken environment: astro dev kill && astro dev start to remove and recreate containers
• Upgrade Airflow runtime by editing the FROM line in Dockerfile and restarting the environment