DevOps Applications in Logistics: Real-World Examples

# .github/workflows/tracking-api.yml
name: tracking-api
on:
  push:
    branches: [main]
jobs:
  test-build:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-python@v5
        with:
          python-version: '3.12'
      - run: pip install -r requirements.txt
      - run: pytest -q
      - run: docker build -t tracking-api:${{ github.sha }} .

#!/usr/bin/env bash
# Check the new version, roll back if errors are high
set -euo pipefail
DEPLOY=tracking-api

kubectl rollout status deployment/$DEPLOY --timeout=120s || {
  echo "rollout did not complete, rolling back"
  kubectl rollout undo deployment/$DEPLOY
  exit 1
}
kubectl rollout history deployment/$DEPLOY | tail -3

STAGES = [
    {"name": "pilot", "sites": ["WH-PUNE-1"], "share": 0.1},
    {"name": "site", "sites": ["WH-PUNE-1"], "share": 1.0},
    {"name": "region", "sites": ["WH-PUNE-1", "WH-NASIK-1"], "share": 1.0},
]

def should_update(device_site: str, device_hash: int, stage: dict) -> bool:
    if device_site not in stage["sites"]:
        return False
    return (device_hash % 100) < stage["share"] * 100

stage = STAGES[0]
for dev, h in [("D1", 4), ("D2", 57), ("D3", 9)]:
    print(dev, should_update("WH-PUNE-1", h, stage))

def go_no_go(crash_rate, avg_scan_ms, open_critical):
    reasons = []
    if crash_rate > 0.01:
        reasons.append(f"crash rate {crash_rate:.1%} above 1%")
    if avg_scan_ms > 800:
        reasons.append(f"scan time {avg_scan_ms}ms above 800ms")
    if open_critical:
        reasons.append(f"{open_critical} critical issue(s) open")
    return ("GO", []) if not reasons else ("NO-GO", reasons)

print(go_no_go(0.004, 520, 0))
print(go_no_go(0.02, 950, 1))

apiVersion: batch/v1
kind: CronJob
metadata:
  name: route-planner
spec:
  schedule: "30 1 * * *"        # 01:30 every night
  concurrencyPolicy: Forbid
  jobTemplate:
    spec:
      backoffLimit: 3             # retry up to 3 times
      activeDeadlineSeconds: 7200
      template:
        spec:
          restartPolicy: Never
          containers:
            - name: planner
              image: example/route-planner:1.4.0
              args: ["--date", "tomorrow"]

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: tracking-api
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: tracking-api
  minReplicas: 4
  maxReplicas: 30
  metrics:
    - type: Resource
      resource:
        name: cpu
        target:
          type: Utilization
          averageUtilization: 60

Project: Parcel tracking platform (dummy data)

Services:
- tracking-api (FastAPI) with tests
- route-planner CronJob that writes a routes file

DevOps:
- CI pipeline building Docker images
- Kubernetes deployment on kind or minikube
- Rollback script and documented drill
- HPA with a k6 load test
- Alert when the CronJob fails

Peak-season readiness

[ ] Load test at expected peak plus margin
[ ] Autoscaling limits reviewed
[ ] Database connection limits checked
[ ] Change freeze dates agreed
[ ] On-call rota published
[ ] Rollback tested in the last two weeks
[ ] Dashboards for tracking, booking, jobs

Runbook: route-planner failed

1. Check last job logs: kubectl logs job/<name>
2. Check input data arrived (orders count > 0)
3. If input missing: contact data team, rerun after fix
4. If code error: roll back image tag, rerun manually
5. Inform dispatch team of expected delay
6. Write a short incident note