Ferhat RudvanoğullarıData Engineer
← All work

Deployment & DevOps

CI/CD & Server Consolidation

Set up automated deploys, a reverse proxy and one-command rollbacks, then moved every app onto a single server.

DevOps

May–Aug 2026

In numbers

Docker build context
1.16 GB → 64 KB
Docker build context
scripts for migration, deploy, rollback and diagnostics
~20
scripts for migration, deploy, rollback and diagnostics
to roll back a release
1 command
to roll back a release

The problem

Apps were being deployed by hand to different servers. If a release went wrong there was no standard way back, and Docker builds were far slower than they needed to be.

What I did

  1. 1

    Automatic deploys

    I set up GitHub Actions with self-hosted runners on our own servers: a push builds the app and ships it to the right environment, with separate pipelines for staging and production.

  2. 2

    Reverse proxy

    Traefik sits in front of everything, routing each request to the right app by domain and handling HTTPS certificates on its own. I added HSTS and other hardening to staging.

  3. 3

    Faster builds

    Every build was shipping dependency folders to Docker along with the source. A proper .dockerignore cut the build context from 1.16 GB to 64 KB, and builds started finishing in seconds.

  4. 4

    One server

    I moved all the apps onto a shared server and wrote scripts for migrating, deploying, rolling back and checking current state. Rolling back became a single command.

Result

Releases and rollbacks take one command, and every new feature is tested in its own isolated environment before it reaches production.

What I learned

A deploy without a rollback plan isn't a deploy, it's a gamble.

Stack

  • Docker
  • Traefik
  • GitHub Actions
  • GCP
  • Linux (systemd)