DevOps Applications in Retail Real-World Examples

# Split stores into rollout waves: pilot first, then regions
stores = {
    "pilot": ["NOI-01", "DEL-04"],
    "north": ["NOI-02", "DEL-01", "DEL-02", "GZB-01"],
    "west": ["MUM-01", "PUN-01", "PUN-02"],
}

def healthy(store):
    # Replace with a real check: recent bills, payment errors, app version
    return True

for wave, store_list in stores.items():
    print(f"Deploying wave '{wave}': {store_list}")
    if not all(healthy(s) for s in store_list):
        print("Health check failed - stopping rollout")
        break
else:
    print("Rollout complete")

# rollout.yml - update POS app on one group of store machines
# run: ansible-playbook -i inventory.ini rollout.yml --limit pilot
- hosts: all
  become: true
  serial: 1                 # one store at a time inside the group
  tasks:
    - name: Install new POS package
      ansible.builtin.apt:
        deb: /tmp/pos-app_5.3.0_amd64.deb
    - name: Restart POS service
      ansible.builtin.systemd:
        name: pos-app
        state: restarted
    - name: Wait for health endpoint
      ansible.builtin.uri:
        url: http://localhost:8080/health
        status_code: 200
      retries: 10
      delay: 6

# Nginx: cache product images and static files at the edge/browser
location ~* \.(jpg|jpeg|png|webp|css|js)$ {
    expires 7d;
    add_header Cache-Control "public, max-age=604800";
}

# Never cache cart or checkout pages
location /checkout {
    add_header Cache-Control "no-store";
    proxy_pass http://checkout_service;
}

# .github/workflows/deploy.yml (excerpt)
# Block production deploys during the sale freeze window
jobs:
  freeze-check:
    runs-on: ubuntu-latest
    steps:
      - name: Stop deploys during sale freeze
        run: |
          today=$(date -u +%F)
          if [[ "$today" > "2026-11-24" && "$today" < "2026-11-30" ]]; then
            echo "Sale freeze is active - deploy blocked"
            exit 1
          fi

import csv, sys

# prices.csv: sku, mrp, selling_price
errors = []
with open("prices.csv", newline="") as f:
    for row in csv.DictReader(f):
        mrp, price = float(row["mrp"]), float(row["selling_price"])
        if price <= 0:
            errors.append(f"{row['sku']}: non-positive price")
        elif price > mrp:
            errors.append(f"{row['sku']}: price above MRP")
        elif price < 0.3 * mrp:
            errors.append(f"{row['sku']}: discount over 70% needs approval")

if errors:
    print("\n".join(errors))
    sys.exit(1)      # fail the pipeline, prices are not published
print("Price file OK")

# Kubernetes: check and undo a bad checkout release
kubectl rollout status deployment/checkout --timeout=120s

# If error alerts fire after the release:
kubectl rollout history deployment/checkout
kubectl rollout undo deployment/checkout

# Confirm the previous version is serving again
kubectl rollout status deployment/checkout

Retail release checklist

[ ] Change reviewed and tested in staging
[ ] Not inside a sale freeze window
[ ] Price/offer files passed validation
[ ] Rollout plan: pilot -> region -> all
[ ] Rollback steps written and tested
[ ] Dashboards ready: orders, payments, errors
[ ] On-call engineer and store support informed
[ ] Store teams told what changes for them

Sale-day runbook (outline)

T-14 days  Load test at expected peak
T-7 days   Raise scaling limits, warm caches
T-3 days   Start change freeze
T-1 day    Final health checks, on-call rota confirmed
Sale day   War room: watch orders, payments, latency
T+1 day    Review incidents and capacity used
T+7 days   Post-sale report and improvements

Incident summary

What happened: ______________________
When detected and how: _______________
Customer / store impact: _____________
Immediate fix: _______________________
Root cause: __________________________
New safeguard added: _________________
Owner and due date: __________________