GCP Radar

GCP Resource Inventory Scanner & Architecture Diagram Generator

A Python CLI tool that inventories all your GCP resources and generates beautiful draw.io architecture diagrams.

Features

• 🔍 Inventory Scanning: Scans Compute Engine, Cloud SQL, Cloud Functions, GKE, Cloud Storage, Cloud Run, Firestore, Pub/Sub, and more
• 💰 Commitment Export: Pull every committed use discount (CUD) across a whole billing account to CSV
• 💵 Billing Export: Daily cost by service (last 90 days) from the Cloud Billing BigQuery export
• 📊 CSV Export: Export complete inventory to CSV for further analysis
• 🏗️ Architecture Diagrams: Auto-generate draw.io .drawio files with logical service connections
• ☁️ Multi-Project Support: Scan multiple GCP projects in one go
• 🔐 Service Account Auth: Use gcloud default credentials or specify service account key

Installation

pip install gcp-radar

Or from source:

git clone https://github.com/yourname/gcp-radar.git
cd gcp-radar
pip install -e .

Quick Start

1. Inventory Your GCP Resources

# Scan default project
gcp-radar inventory --export inventory.csv

# Scan specific project
gcp-radar inventory --project my-project --export inventory.csv

# Export to specific region
gcp-radar inventory --region us-central1 --export inventory.csv

2. Export Committed Use Discounts (CUDs)

# Every commitment on the billing account funding your current project
gcp-radar commitments --export commitments.csv

# Pick the billing account explicitly
gcp-radar commitments --billing-account 01ABCD-EF2345-6789GH --export commitments.csv

# Single project, or a single region
gcp-radar commitments --project my-project --export commitments.csv
gcp-radar commitments --region us-central1

# Or fold commitments into a regular inventory scan
gcp-radar inventory --commitments --export inventory.csv

Without --project, gcp-radar resolves the billing account, lists every project attached to it, and scans each one — so you get all commitments on the account, not just the current project's.

3. Export Daily Cost by Service

Reads the Cloud Billing BigQuery export (gcp_billing_export_v1_*), prints the top services, and writes a CSV.

# Standalone: last 90 days, before credits
gcp-radar billing --table my-proj.billing.gcp_billing_export_v1_XXXXXX_XXXXXX_XXXXXX

# Net of credits, 30 days, one project
gcp-radar billing --table ... --days 30 --metric net --project my-project --output cost.csv

# Alongside an inventory scan or one-shot run
gcp-radar inventory --export inventory.csv --billing --billing-table ...
gcp-radar run --output inventory.csv --diagram architecture.drawio --billing-table ...

Set GCP_RADAR_BILLING_TABLE to skip --billing-table every time.

With --billing on inventory/run, --project also limits the billing query.

4. One-Shot: Everything in One CSV

gcp-radar run --project my-project --billing-table my-proj.billing.gcp_billing_export_v1_XXXXXX \
  --output inventory.csv --diagram architecture.drawio   # --diagram is optional

Inventory, CUDs and billing land in a single file. Every row has a RecordType:

--output-csv / --output-diagram still work as aliases for --output / --diagram. Without a billing table (--billing-table or GCP_RADAR_BILLING_TABLE) the billing rows are skipped.

5. Generate Architecture Diagram

gcp-radar diagram --input inventory.csv --output architecture.drawio

# Then open architecture.drawio in app.diagrams.net or the desktop app

Supported GCP Services

Requirements

• Python 3.9+
• GCP credentials (via gcloud auth or service account key)
• Permissions: compute:Read, sql:Read, container:Read, storage:Read, pubsub:Read, firestore:Read, etc.
• For commitments: compute.commitments.list on each project, plus billing.resourceAssociations.list on the billing account (to enumerate its projects) — e.g. roles/billing.viewer
• For billing / --billing: a billing export to BigQuery and bigquery.jobs.create plus bigquery.tables.getData on the export table (e.g. roles/bigquery.jobUser + roles/bigquery.dataViewer)

Output Format

CSV Inventory

CSV Commitments

CSV Billing

Architecture Diagram

Auto-generated .drawio file with:

• Swim lanes by service category (Compute, Database, Storage, Messaging)
• Icons for each resource
• Logical connections between services

Configuration

Create a .env file to override defaults:

export GCP_PROJECT=my-project
export GCP_REGION=us-central1

Or pass flags:

gcp-radar inventory --project my-project --region us-central1

License

MIT

Contributing

Pull requests welcome! Please open an issue first to discuss changes.

Support

GitHub Issues: https://github.com/yourname/gcp-radar/issues