gcloud_skill

Overview

This skill is a practical guide for implementing and using the Google Cloud SDK (gcloud CLI) to manage Google Cloud resources, automate operations, and integrate cloud tasks into scripts and CI/CD pipelines. It focuses on installation, authentication, configuration management, common workflows (Compute/GKE/Cloud Run/App Engine/Storage), and reliable automation patterns. The content emphasizes secure practices like service account impersonation and idempotent scripting.

How this skill works

The skill explains gcloud command structure, release levels (alpha/beta/GA), components, and global flags, then walks through installation methods for Linux, macOS, Windows, and package-managed installs. It covers authentication options (user accounts, service accounts, ADC, impersonation), named configurations for multi-environment setups, and examples for resource operations. Scripting sections show output formats, filtering, error handling, retries, and CI/CD integration patterns.

When to use it

• Installing or upgrading the Google Cloud SDK on developer machines or CI runners
• Authenticating to Google Cloud with user accounts, service accounts, or ADC
• Managing multiple projects, regions, or environment-specific configurations
• Deploying and managing Compute Engine, GKE, Cloud Run, App Engine, or Cloud Storage
• Automating gcloud operations in scripts and CI/CD pipelines
• Troubleshooting authentication, permissions, or deployment failures

Best practices

• Prefer service account impersonation in production to avoid long-lived keys
• Use named gcloud configurations for dev/staging/prod to avoid accidental changes
• Format output as JSON for reliable parsing in scripts and CI jobs
• Apply server-side filtering with --filter to reduce API and parsing load
• Make scripts idempotent and include retry/exponential backoff for transient errors
• Limit scope of IAM roles; grant least privilege and audit bindings regularly

Example use cases

• Set up a developer laptop: install gcloud, run gcloud init, and create a dev configuration
• CI pipeline: activate service account with key file or use workload identity, set project, and deploy a Cloud Run service
• Operational task: list running Compute Engine instances filtered by zone and status, then stop stale instances
• Kubernetes: create a GKE cluster, fetch credentials with get-credentials, and run kubectl commands
• Backup flow: sync local directory to Cloud Storage using gsutil rsync in a scheduled job