Introduction to Cloud Composer

A Managed Airflow Service on Google Cloud

Pradeep Singh
3 min readMar 6, 2023
Photo by Devashish at googlecloudcommunity.com

This is the first article of the series ‘Mastering Cloud Composer: A Comprehensive Guide to Managed Airflow on Google Cloud’.

In this article, we will go through the below topics:

  • Overview of Cloud Composer and its features
  • Comparison with Airflow deployment options (e.g., self-managed, hosted)
  • Benefits of using Cloud Composer for workflow management

Overview of Cloud Composer

Cloud Composer is a fully managed workflow orchestration service that is built on top of the open-source Apache Airflow project. It is designed to simplify the management and deployment of complex workflows by providing an easy-to-use interface for creating and scheduling workflows. With Cloud Composer, users can build workflows using the Airflow API and Web UI, manage dependencies, and monitor the execution of workflows.

TLDR: Cloud Composer is nothing but Apache Airflow entirely managed by Google for you, start creating your workflows as per your need.

Features of Cloud Composer

Cloud Composer provides several features that make it an attractive option for workflow management. These features include:

  1. Managed Service: Cloud Composer is a fully managed service that handles the deployment, scaling, and monitoring of Airflow clusters, allowing users to focus on creating workflows.
  2. Integration with GCP Services: Cloud Composer integrates with other Google Cloud Platform (GCP) services such as BigQuery, Cloud Storage, and Dataflow, making it easy to incorporate these services into workflows.
  3. Versioning and Rollback: Cloud Composer allows users to version their workflows and roll back to previous versions if needed.
  4. Secure by Default: Cloud Composer uses Google Cloud IAM to manage access to workflows, ensuring that only authorized users can access sensitive data.

Self vs GCP Managed Airflow

There are two ways to manage Airflow workflows: self-hosting, where you manage everything yourself, or using a managed service like Google Cloud Composer, where Google manages everything for you.:

  1. Self-managed Airflow: Self-managed Airflow requires users to deploy and manage their own Airflow clusters. This option provides more flexibility and control but requires more expertise in managing Airflow clusters.
  2. Managed Airflow on GCP: Cloud Composer provides a more integrated and seamless workflow management experience by integrating with other GCP services such as BigQuery, Cloud Storage, and Dataflow. It also has a simpler pricing model, with users only paying for the resources used by their Airflow environment. Additionally, Cloud Composer is fully managed, reducing the management overhead required by users compared to self-managed Airflow.

Benefits of using Cloud Composer for workflow management

Using Cloud Composer for workflow management provides several benefits to users, including:

  1. Reduced Management Overhead: Cloud Composer is a fully managed service that handles the deployment, scaling, and monitoring of Airflow clusters, reducing the management overhead required by users.
  2. Integration with GCP Services: Cloud Composer integrates with other GCP services, making it easy to incorporate these services into workflows.
  3. Easy-to-Use Interface: Cloud Composer provides an easy-to-use interface for creating, scheduling, and monitoring workflows, reducing the learning curve required for users to get started.

Components of Cloud Composer

The following are the main components of Cloud Composer:

Web server: The GUI to track the status of jobs. Stores metadata related to DAGs, jobs, etc. Google Cloud uses as a Web Server. In Google Cloud, the Airflow Web Server is hosted on Compute Engine instance and allows you to manage access to it using the IAM policy.

Scheduler: Responsible for orchestrating and scheduling jobs. In Google Cloud, the Airflow Scheduler is used and is implemented as a Kubernetes Deployment.

Executor: Set of worker processes that are responsible for executing the tasks in the workflow. In Google Cloud, the Airflow Executor is used and is implemented as a Kubernetes Pod.

Metadata database: Stores metadata related to DAGs, jobs, etc. Google Cloud uses Cloud SQL as a Metadata database, which provides high availability, automatic backups, and automated security patching.

Additionally, DAGs, Logs, and Plugins are stored in Cloud Storage.

Cloud Composer is a managed workflow orchestration service that provides an easy-to-use interface for creating, scheduling, and monitoring workflows. It integrates with other GCP services, providing a seamless workflow management experience for users. Compared to other Airflow deployment options, Cloud Composer reduces the management overhead required by users and provides a more integrated workflow management solution.

--

--

Pradeep Singh

MLOps Engineer @ Genpact / psrajput.com / Running (10k in 59.12, 5k in 26.15) / Cricket / Trekking / Chess