Deploying a Site on Selectel: Terraform, Kubernetes, CI/CD

You've deployed a site on Selectel, but a week later it goes down due to incorrect security group configuration? Or deployment takes an hour, and rollback is a full operation? We solve these problems with automated deployment using Terraform, Kubernetes, and GitHub Actions. Rollouts take minutes, infrastructure is reliable, and it complies with 152-FZ. Once a project went down due to incorrect Object Storage configuration — images didn't load, LCP rose to 5 seconds. We rewrote the deployment with Terraform, set up CDN, and LCP dropped to 0.8 seconds. This is just one example of how automation saves business.

Our team has 5+ years of experience with Selectel infrastructure and has set up deployment for 30+ projects, from landing pages to high-load SaaS platforms. Here's how we do it.

Problems We Solve

When deploying on Selectel, we frequently encounter typical errors:

• N+1 queries when working with Object Storage — incorrect CDN configuration leads to delays. We fix this through content preloading and proper Cache-Control headers.
• Hydration mismatch in React/Next.js when using SSR — due to different server and client response times. We solve it by configuring synchronization via Edge Functions.
• Data leaks due to incorrect security group configuration. We always use minimal privileges: only 80/443 and SSH from whitelisted IPs.

Each problem is accompanied by a concrete solution we've refined on dozens of projects.

How to Organize Deployment for 152-FZ Compliance

Selectel provides certified infrastructure. We configure:

• Encryption at rest via OpenStack Barbican.
• Access logging to Object Storage via S3 Access Logs.
• Regular database backups to a separate project.

This ensures compliance without extra costs.

Why Selectel Is More Profitable Than Foreign Clouds

Selectel is 2–3 times cheaper than AWS/Azure for Russian projects when considering latency and compliance. For example, using Managed Kubernetes on Selectel reduces administration costs by 40% compared to manual deployment. Plus, data stays in Russia, which is critical for 152-FZ.

How to Set Up Deployment on Selectel from Scratch

We use a proven stack: Terraform 1.7 for infrastructure description, Ansible for configuration, GitHub Actions for CI/CD. One project — an online store on Laravel 11: we terraformed the project, set up Object Storage for media files, and Managed Kubernetes for the frontend. Deployment takes 3 minutes, rollback is 1 command.

# Creating a server via CLI pip install python-openstackclient export OS_AUTH_URL=https://cloud.api.selcloud.ru/identity/v3 export OS_PROJECT_ID=your_project_id export OS_USERNAME=your_username export OS_PASSWORD=your_password export OS_REGION_NAME=ru-1 openstack server create \ --flavor 1014 \ --image "Ubuntu 22.04 LTS 64-bit" \ --key-name my-key \ --security-group web-sg \ myapp-prod 

# Terraform for Selectel terraform { required_providers { selectel = { source = "selectel/selectel" version = "~> 5.0" } openstack = { source = "terraform-provider-openstack/openstack" version = "~> 1.53" } } } provider "selectel" { token = var.selectel_token } resource "selectel_vpc_project_v2" "myapp" { name = "myapp-project" theme = { color = "#4CAF50" } } resource "openstack_compute_instance_v2" "app" { name = "myapp-prod" flavor_name = "1014" image_name = "Ubuntu 22.04 LTS 64-bit" key_pair = openstack_compute_keypair_v2.mykey.name security_groups = ["web-sg", "ssh-sg"] region = "ru-1" network { name = "public" } user_data = file("cloud-init.yaml") } 

# Selectel Object Storage (Swift/S3) aws configure aws --endpoint-url https://s3.selcdn.ru s3 mb s3://myapp-assets aws --endpoint-url https://s3.selcdn.ru \ s3 sync ./dist/assets s3://myapp-assets \ --acl public-read \ --cache-control "public, max-age=31536000, immutable" 

# Managed Kubernetes (MKS) selectel mks kubeconfig --cluster-id CLUSTER_ID > kubeconfig.yaml export KUBECONFIG=./kubeconfig.yaml kubectl apply -f k8s/ kubectl get pods -n myapp 

# GitHub Actions for Selectel VPS jobs: deploy: steps: - name: Deploy uses: appleboy/ssh-action@v1 with: host: ${{ secrets.SELECTEL_SERVER_IP }} username: deploy key: ${{ secrets.SSH_KEY }} script: | cd /var/www/myapp git pull origin main composer install --no-dev --optimize-autoloader npm ci && npm run build php artisan migrate --force php artisan optimize sudo systemctl reload php8.3-fpm nginx 

Typical Deployment Mistakes and Their Prevention

• Ignoring security groups — leaving SSH open to all IPs. We only use whitelists.
• Incorrect CORS configuration for S3 — static files don't load on the frontend. We explicitly specify allowed origins.
• Lack of monitoring — finding out about downtime from clients. We set up alerts via SendGrid or Telegram.

These mistakes are easy to avoid if you follow our practices.

Process

1. Analysis — determine requirements for load, budget, compliance.
2. Design — draw architecture: VPS vs Kubernetes, S3 vs DB, CDN.
3. Infrastructure — deploy project, networks, security groups with Terraform code.
4. Implementation — configure application, CI/CD, monitoring (Prometheus + Grafana).
5. Testing — load testing, Core Web Vitals check.
6. Deployment — canary rollout with automatic rollback on errors.

What's Included in Deployment Setup

• Architecture design for your tasks.
• Writing Terraform configurations for the entire infrastructure.
• Setting up CI/CD (GitHub Actions or GitLab CI) with automatic deployment and rollback.
• Configuring Object Storage and CDN for static files.
• Integrating monitoring (Prometheus + Grafana) with alerts.
• Documentation on deployment and training for your team.
• Technical support after launch.

Comparison: VPS vs Managed Kubernetes

Deployment Component Timelines

The cost is calculated individually per project. Request a custom deployment cost estimate — we'll select the optimal architecture. Get a consultation — we'll assess your project in one business day. Contact us to set up a deployment that won't let you down.