MathOps Microservice

Backend-only FastAPI service that queues math operations to a Celery worker via RabbitMQ and stores results in PostgreSQL.

Tech stack

• Python 3.11
  • FastAPI async API
  • Celery worker (separate process/container)
  • PostgreSQL persistence
  • Alembic migrations
• Docker-ready
• Minikube-compliant

Project layout

app/
  api/         # HTTP routes
  core/        # config + exception handlers
  db/          # session + base
  models/      # SQLAlchemy models
  schemas/     # Pydantic schemas
  services/    # business logic
  tasks/       # Celery app + tasks

Local development (docker-compose)

1. Create a local .env file:

cp .env.example .env

2. Start the stack:

docker compose up --build

3. Run migrations (external tool, not runtime-managed):

docker compose exec api alembic upgrade head

4. Open the API docs:

http://localhost:8000/docs

Adminer is available at http://localhost:8080 for safe DB access (no direct DB port exposure).

Local development (no Docker)

You need RabbitMQ and PostgreSQL running locally. Then:

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

alembic upgrade head
python -m app.server

In another terminal:

celery -A app.tasks.celery_app:celery_app worker --loglevel=info

Docs: http://localhost:8000/docs.

Build and deploy on Windows

Prerequisites:

• Docker, Minikube, Kubectl installed

Run:

runDeploy.bat

The script will build the image, apply Kubernetes resources, run migrations, and open /docs.

Build and deploy on Linux distros

Prerequisites:

• Docker, Minikube, Kubectl installed

Run:

./UbuntuDeploy.sh

The script will build the image, apply Kubernetes resources, run migrations, and open /docs.

Kubernetes notes

• math-deployment.yaml includes the API, worker, RabbitMQ, and PostgreSQL deployments.
• Configuration is driven via math-config.yaml (ConfigMap).
• The API service is exposed via NodePort; open /docs for the Swagger UI.
• Migrations are run via the math-migrate-job.yaml Job.

Kubernetes workflow (manual)

minikube start
docker build -t mathcontainer:latest .
minikube image load mathcontainer:latest
kubectl apply -f math-config.yaml -f math-deployment.yaml
kubectl rollout status deployment/math-api --timeout=120s
kubectl delete job math-migrate --ignore-not-found
kubectl apply -f math-migrate-job.yaml
kubectl wait --for=condition=complete job/math-migrate --timeout=120s
minikube service math-calculator-service --url

Endpoints

• POST /calculate -> returns a job ID (string)
• GET /result/{job_id} -> returns { status, result }

Database access tools

Use DBeaver or Adminer. Adminer via container is safer; exposing the DB port directly is not recommended.