The code we build in the Day 1 live session. One idea runs through all of it:

You write plain Python functions that do DevOps work, and those same functions get a second front door — an HTTP API — without changing the logic.

• aws_functions.py — boto3 helpers (list_buckets, s3_summary, ec2_summary, find_public_buckets) plus a main() that prints a report and writes aws_report.json. With no AWS credentials it returns labelled demo data, so it runs on any laptop.
• metrics.py — get_system_metrics() using psutil.
• logs.py — analyze_logs() counting INFO / WARNING / ERROR in a log file.
• hello_api.py — the smallest FastAPI app, to learn the basics.
• api.py — the DevOps Utilities API. It imports the functions above and exposes them as endpoints. This is the "second front door".
• app.log — sample log (INFO=10, WARNING=2, ERROR=3).
• Dockerfile — ship the API as a container.

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

The /aws/* endpoints and the report work without an AWS account — with no credentials they return labelled demo data (look for "demo": true). Configure credentials when you want real results:

aws configure   # or export AWS_ACCESS_KEY_ID / AWS_SECRET_ACCESS_KEY / AWS_DEFAULT_REGION

python aws_functions.py     # prints an S3 + EC2 report, writes aws_report.json
python metrics.py           # CPU / memory / disk
python logs.py              # log-level counts

Everything here reads from AWS only. Nothing is created, modified, or deleted.

Warm up with the tiny app:

uvicorn hello_api:app --reload
# open http://127.0.0.1:8000/docs

Then the utilities API:

uvicorn api:app --reload

curl http://127.0.0.1:8000/health
curl http://127.0.0.1:8000/metrics
curl "http://127.0.0.1:8000/metrics?cpu_threshold=50"
curl http://127.0.0.1:8000/logs
curl http://127.0.0.1:8000/aws/s3              # buckets, new vs old
curl http://127.0.0.1:8000/aws/ec2             # instances + state
curl http://127.0.0.1:8000/aws/public-buckets  # security check: which buckets are public

The /aws/* endpoints return demo data when no credentials are set, so they respond even without an AWS account.

Interactive docs: http://127.0.0.1:8000/docs

docker build -t devops-utilities-api .
docker run -p 8000:8000 devops-utilities-api
curl http://127.0.0.1:8000/health

1. Add a /version endpoint returning {"version": "1.0.0"}.
2. Add /logs/errors returning only the ERROR count (reuse analyze_logs).
3. Add a /aws/buckets endpoint returning just bucket names (reuse list_buckets).
4. Bonus: build the Docker image and hit /health.

This flat layout keeps the live session simple. A larger project would split the logic into services/ and the HTTP layer into routers/ — see 07-apis-with-fastapi/devops-utilities-api/.