ACEest Fitness & Gym — DevOps CI/CD

Student: Atul Yadav
Roll Number: 2024tm93580
Email: 2024tm93580@wilp.bits-pilani.ac.in
Course: Introduction to DEVOPS (CSIZG514 / SEZG514) — S1-25 · Assignment 2
GitHub: https://github.com/atul411/devops-assignment-2

A Flask-based fitness management application with a complete CI/CD pipeline implementation for the Introduction to DEVOPS (CSIZG514/SEZG514) Assignment 2.

The original assignment provided a Tkinter desktop app across 10 incremental versions; this repository ports the domain (programs, clients, workouts, progress) into a Flask web service and wraps it in a fully automated DevOps pipeline.

Quick start

# Local dev
make install
make test           # 48 pytest tests, ~80% coverage
make run            # http://localhost:5000

# Containerised
docker-compose up -d --build

# Kubernetes (Minikube)
kubectl apply -k k8s/base/                 # baseline (Rolling Update default)
kubectl apply -k k8s/blue-green/           # OR Blue-Green
kubectl apply -k k8s/canary/               # OR Canary
kubectl apply -k k8s/shadow/               # OR Shadow (Istio)
kubectl apply -k k8s/ab-testing/           # OR A/B Testing

Application versions

The same Docker image runs as v1 / v2 / v3 by toggling FEATURE_LEVEL (1, 2, or 3):

Level	Routes registered	Mirrors Tkinter
1	/programs (read-only program info)	Aceestver-1.0.py
2	+ /clients CRUD + progress tracking	Aceestver-2.x.py
3	+ /login, /workouts, /metrics, /program-generator, /dashboard	Aceestver-3.x.py

Default admin user: admin / admin (password is hashed with PBKDF2-SHA256).

API surface (FEATURE_LEVEL=3)

Method	Path	Notes
GET	/health	Liveness — always 200
GET	/ready	Readiness — pings SQLite
GET	/programs	HTML or ?format=json
GET	/programs/<key>	FL, MG, BG
GET	/programs/<key>/calories?weight=70	Calorie calc (factor × weight)
GET / POST	/clients/	List / create
GET / PATCH / DELETE	/clients/<name>	Detail / update / delete
GET / POST	/clients/<name>/progress	Weekly adherence log
POST	/login / /logout	Session-based auth
GET / POST	/workouts/<name>	Workouts per client
POST	/workouts/<id>/exercises	Exercise sets/reps/weight
GET / POST	/metrics/<name>	Body weight / waist / bodyfat
POST	/program-generator/<name>	Random AI-style program

DevOps pipeline

Developer push -> GitHub
         |
         v
  Jenkins (pollSCM 1m) ---> Checkout -> Setup -> Lint (flake8) -> Pytest (--cov)
                                                                   |
                                                                   v
                                                 SonarQube scan -> Quality Gate (fail = abort)
                                                                                |
                                                                                v
                                                            docker build -> docker push
                                                                                |
                                                                                v
                                                               kubectl apply -k k8s/<strategy>/
                                                                                |
                                                                                v
                                                                  Deployment health check
                                                                  (auto-rollback on failure)

See docs/architecture.md for the full diagram and docs/REPORT.md for the 2-page report.

Repository layout

app/                Flask application (factory + 5 blueprints)
tests/              Pytest suite (48 tests, ≥80% coverage)
k8s/                Kubernetes manifests
  base/             Namespace + ConfigMap + Secret + PVC + Deployment + Service + Ingress
  rolling-update/   Default RollingUpdate (maxSurge=1, maxUnavailable=0)
  blue-green/       Two Deployments + Service selector switch
  canary/           Stable + canary, nginx-ingress canary-weight
  shadow/           Istio VirtualService.mirror (+ nginx fallback)
  ab-testing/       Two variants, header-routed via X-Variant: B
scripts/            bluegreen-switch.sh, canary-promote.sh, rollback.sh
docs/               REPORT.md (assignment write-up), architecture.md
legacy-tkinter/     Original Tkinter sources preserved for reference
Dockerfile          Multi-stage, python:3.12-slim, gunicorn, non-root, healthcheck
Jenkinsfile         Declarative pipeline with SonarQube quality gate
sonar-project.properties
.github/workflows/  Optional GitHub Actions secondary CI
docker-compose.yml  Local dev with persistent SQLite volume
Makefile            Common dev tasks

Configured identifiers

This repo is pre-configured for the following identities — replace if you fork:

Identifier	Value	Used in
GitHub repo	atul411/devops-assignment-2	Dockerfile LABEL, README links
Docker Hub repo	atulyadav123/aceest-fitness	Dockerfile, Jenkinsfile, all k8s/*/*.yaml
Jenkins credential ID	dockerhub-credentials	Jenkinsfile (configure under Jenkins → Credentials)
Jenkins credential ID	kubeconfig	Jenkinsfile (kubeconfig file credential)
Ingress hostname	aceest.local	All k8s/**/ingress*.yaml (replace with your real hostname)

Submission checklist

• Flask app + 3 versions (single image, FEATURE_LEVEL toggle)
• Pytest suite (48 tests, 80.9% coverage)
• Dockerfile (multi-stage, non-root, healthcheck) — verified via GitHub Actions
• Jenkinsfile (declarative, pollSCM, SonarQube quality gate, auto-rollback)
• Kubernetes manifests for all 5 deployment strategies — validated against live API
• SonarQube analysis executed — Quality Gate PASSED (0 bugs, 0 vulns, 0 smells, all A ratings)
• Rollback scripts (bluegreen-switch, canary-promote, rollback)
• 2-3 page report (docs/REPORT.md)
• Submission checklist (docs/SUBMISSION.md)
• Deployment playbook (docs/DEPLOYMENT.md)
• GitHub repo public at https://github.com/atul411/devops-assignment-2
• GitHub Actions runs green
• Docker Hub push (needs your DOCKERHUB_TOKEN secret — workflow auto-pushes once set)
• Public cluster endpoint (needs cloud K8s — see DEPLOYMENT.md § 4(B))
• Live Jenkins build screenshot (Jenkins is running locally; complete plugin install via UI)