Google Cloud Cheatsheet
Services Overview
Use this Google Cloud reference while you build software engineering projects, review code, or refresh the syntax you reach for most.
Core Compute
| Service | What it is | Best for |
|---|---|---|
| Compute Engine | IaaS VMs (any OS, full control) | Lift-and-shift, custom runtimes |
| GKE | Managed Kubernetes | Containerized microservices at scale |
| Cloud Run | Serverless containers (HTTP) | Stateless APIs, event-driven services |
| Cloud Functions | Serverless functions (event triggers) | Lightweight glue, webhooks |
| App Engine | Managed PaaS (standard/flexible) | Web apps with zero infra ops |
| Batch | Managed batch job scheduling | HPC, large ML training jobs |
Core Storage & Databases
| Service | Type | Best for |
|---|---|---|
| Cloud Storage | Object / blob | Files, backups, static assets, data lake |
| Persistent Disk | Block storage | VM boot disks, databases |
| Filestore | Managed NFS | Shared file systems across VMs/GKE |
| Cloud SQL | Managed relational (PostgreSQL, MySQL, SQL Server) | OLTP workloads |
| Cloud Spanner | Globally distributed SQL | Planet-scale ACID transactions |
| Firestore | Serverless NoSQL document DB | Mobile/web real-time apps |
| Bigtable | Managed wide-column NoSQL | Time-series, IoT at massive scale |
| Memorystore | Managed Redis / Memcached | Caching, session storage |
Data & Analytics
| Service | What it does |
|---|---|
| BigQuery | Serverless data warehouse — SQL analytics over petabytes |
| Dataflow | Managed Apache Beam (streaming + batch ETL) |
| Pub/Sub | Async messaging / event streaming |
| Dataproc | Managed Spark + Hadoop |
| Looker Studio | BI dashboards on top of BigQuery |
Networking
| Service | What it does |
|---|---|
| VPC | Private network, subnets, firewall rules |
| Cloud Load Balancing | Global / regional HTTP(S), TCP, UDP LB |
| Cloud CDN | Edge caching for GCS/backend content |
| Cloud DNS | Managed authoritative DNS |
| Cloud NAT | Outbound NAT for private VMs |
| Cloud Interconnect / VPN | Private connectivity to on-prem |
AI & ML
| Service | What it does |
|---|---|
| Vertex AI | Unified ML platform — training, serving, pipelines |
| Gemini API | Access Google's Gemini models |
| Vision / NL / Speech APIs | Pre-trained models via REST |
| AutoML | Train custom models with no ML code |
DevOps & Security
| Service | What it does |
|---|---|
| Cloud Build | CI/CD — build, test, deploy pipelines |
| Artifact Registry | Docker images, language packages |
| Secret Manager | Store and audit API keys/secrets |
| IAM | Identity, roles, and permissions |
| Cloud Armor | WAF + DDoS protection |
| Binary Authorization | Deploy only signed container images |
Global Footprint
- Regions — 40+ geographic regions (e.g.,
us-central1,europe-west1,asia-east1) - Zones — 1+ isolated data-center zones per region (e.g.,
us-central1-a) - Multi-region — storage class spanning 2+ regions (
US,EU,ASIA)
Pick a region close to your users. For high availability, deploy across two zones minimum in the same region or use global load balancing across regions.
Resource Hierarchy
Organization └── Folders (optional grouping) └── Projects ← billing, APIs, IAM boundary └── Resources (VMs, buckets, functions…)
- Project is the fundamental unit — all resources live in a project.
- IAM policies and billing attach at Organization, Folder, or Project level (inherited downward).
- Enable APIs per project:
gcloud services enable compute.googleapis.com
Pricing Model
| Concept | How it works |
|---|---|
| Pay-as-you-go | Billed per second / per byte / per request |
| Committed Use Discounts | 1- or 3-year commits for Compute/SQL → up to 57% off |
| Sustained Use Discounts | Auto-applied for VMs running >25% of month |
| Spot / Preemptible VMs | Up to 91% cheaper; can be reclaimed anytime |
| Free Tier | Always-free resources (see the Cost and Free Tier topic) |