Why should you care?

Lift-and-shift VMs miss autoscale, managed DB, and pay-per-use economics.

See it live — copy this example

Paste into your MeanVerse project (Angular + Express + MongoDB), then run with ng serve / node / mongosh as noted.

# Twelve-factor style config
# III. Config in env / secrets manager
# VI. Stateless processes — session in Redis not disk
# IX. Disposability — graceful SIGTERM handler
process.on('SIGTERM', async () => {
  server.close();
  await mongoose.connection.close();
  process.exit(0);
});

What happened?

• Stateless API pods scale horizontally.
• SIGTERM handler drains connections before K8s kills pod.
• Config externalized — same image all environments.

Practice next

1. Audit app for local disk state — move to Redis/S3.
2. build health/readiness endpoints.
3. Use managed Mongo Atlas not self-patched VM.
4. Add liveness vs readiness different checks.
5. Migrate cron jobs to K8s CronJob resources.

Remember

Cloud-native = containers + managed services + observability. Stateless apps scale; state in DB/cache. Graceful shutdown respects orchestrator.

MeanVerse platform upgrade

Company moves from VM monolith to EKS + Atlas + CloudFront.

Outcome: Deploy frequency 1/week to 5/day; infra cost per tenant drops.