Skip to content

Lab — Python GCP Inventory

Lab Overview

Purpose: Produce a GCE instance inventory (name, zone, status, labels).

Scenario: Learners may lack GCP projects; fixtures keep the lab runnable everywhere.

Expected outcome: --fixture path always works; live path optional with Application Default Credentials.

This is a lab, not a tutorial

Apply Cloud Automation — AWS, Azure, GCP.

Business Scenario

A multi-project landing zone needs a weekly inventory of running VMs missing owner labels.

Learning Objectives

  • Flatten instance records to a common schema
  • Flag missing required labels
  • Fixture-first; graceful live failures

Prerequisites

Knowledge

Software

Fixtures required; google-cloud-compute optional. Estimated cost: £0.

Architecture

Cloud inventory fixture and dry-run path

Environment

mkdir -p ~/rebash-lab-python-gcp/{fixtures,out}
cd ~/rebash-lab-python-gcp
python3 -m venv .venv && source .venv/bin/activate

Initial State

cat > fixtures/instances.json << 'EOF'
[
  {"name": "api-1", "zone": "europe-west2-a", "status": "RUNNING", "labels": {"owner": "platform"}},
  {"name": "scratch", "zone": "europe-west2-b", "status": "TERMINATED", "labels": {}}
]
EOF

Task

gcp_inventory.py --fixture fixtures/instances.json → out/inventory.json plus out/missing_labels.json for instances lacking owner.

Validation

python gcp_inventory.py --fixture fixtures/instances.json
python -c 'import json; assert len(json.load(open("out/missing_labels.json")))==1'
  • Fixture works offline
  • Missing label detection works
  • No service account JSON committed

Troubleshooting

Symptom Fix
ADC errors Expected without GCP — use fixtures

Cleanup

deactivate 2>/dev/null || true
rm -rf ~/rebash-lab-python-gcp

Production Discussion

Use workload identity federation for CI; prefer Cloud Asset Inventory for organisation-scale queries.