All cases | Portfolio

AWS → Another Cloud Migration

Cloud Infrastructure Migration from AWS to a Cost-Optimized Private Cloud Client Confidential Challenge The client needed to migrate their entire AWS infrastructure to an alternative cloud provider to reduce costs and eliminate vendor lock-in. The key requirements were minimal downtime, zero functionality loss, and a reliable rollback strategy. Solution 1. Audit & Planning Full inventory of AWS resources (EC2, RDS, S3, VPC) Service mapping: AWS → target cloud equivalents Phased migration plan with clear milestones Rollback strategy for each stage 2. Infrastructure Preparation Terraform for IaC on the target platform VPC, subnets, and security groups configuration Managed PostgreSQL and Redis deployment Object Storage setup (S3-compatible) 3. Data Migration Database replication via DMS S3 → Object Storage sync Docker image transfer to private Container Registry Full validation on staging environment 4. Production Cutover DNS failover for gradual traffic switching Real-time metrics monitoring during cutover AWS kept on standby for emergency rollback Final cutover completed in 2 hours Technologies AWS Terraform Prometheus Grafana GitLab CI Results ✅ Cost reduction: 40% savings (from $15k to $9k/month) ✅ Downtime: only 2 hours vs. 8 hours planned ✅ Vendor independence: full exit from AWS with no functionality loss ✅ Performance: maintained at the same level post-migration ✅ Compliance: data residency and security requirements met

Read

CI/CD & Production Infrastructure for a Social App

Production Infrastructure & CI/CD for a Social App Launch Client Puzzle Master — a social matching platform Challenge The startup had a production-ready Nest.js backend and Angular frontend, but zero infrastructure: deployments were manual, there was no CI/CD, no monitoring, no backups, and no separation between dev and prod environments. The goal was to build a complete DevOps stack from scratch before the public launch. Solution 1. Application Containerization Multi-stage Dockerfile for backend (Nest.js + Prisma, non-root user) Multi-stage Dockerfile for frontend (Angular 12, legacy OpenSSL, Nginx for static assets) Docker Compose full stack: PostgreSQL 15, Redis 7, imgproxy, Nginx Healthchecks and depends_on for correct startup ordering Isolated dev and prod environments in /opt/dev and /opt/prod 2. GitLab CI/CD Migration of repository from Bitbucket to GitLab Pipeline for backend and frontend: build → push → deploy GitLab Container Registry for Docker image storage Automatic deploy to dev on every push; manual trigger for prod SSH deployment to VPS via SSH_PRIVATE_KEY 3. Nginx Reverse Proxy Environment-agnostic config via envsubst for dev/prod parity SSL/TLS (TLSv1.2, TLSv1.3) with Cloudflare certificates Routing: /api/* → backend:4000, /* → frontend:80 www → root domain redirect (301) Separate imgproxy stack with SSL termination 4. Security (Ansible) Server hardening via Ansible: SSH key-only auth, root login disabled UFW Firewall: only ports 80, 443, and custom SSH open Database accessible only via SSH tunnel All secrets stored in GitLab CI/CD variables 5. Monitoring Prometheus + Grafana with automated dashboard provisioning Exporters: Node, cAdvisor, Postgres, Redis, Nginx, Blackbox 5 Grafana dashboards: server, Docker containers, PostgreSQL, Redis, Nginx Alertmanager with Slack/webhook integration; alerts on CPU/RAM/Disk/API/SSL 6. Database Backups Automated pg_dump every hour gzip compression and upload to S3-compatible object storage (Cloudflare R2) Prometheus backup metrics: success status, size, timestamp Alerts: DatabaseBackupMissing, DatabaseBackupFailed, DatabaseBackupSizeAnomaly Technologies GitLab CI Docker Ansible Prometheus Nginx PostgreSQL Results ✅ Deploy: git push to main → automatic build and deploy to server ✅ Environments: full dev/prod isolation on a single VPS ✅ Monitoring: 5 dashboards, alerts across 6 categories ✅ Backups: automated hourly pg_dump to Cloudflare R2 ✅ Security: UFW, key-based SSH, database inaccessible from outside ✅ Scalability: architecture ready for database extraction to a dedicated server

Read

CI/CD with GitLab + Kubernetes: deploy in 10 minutes

CI/CD automation: from 2 hours to 10 minutes per deploy Client E-commerce startup, 5-person engineering team Challenge Manual deployment took 2 hours Frequent release errors No fast rollback path Required: CI/CD automation, GitOps, fast rollback Solution Self-hosted GitLab GitLab CI pipeline (Build → Test → Deploy) Managed Kubernetes in Yandex Cloud Flux CD for GitOps Technologies GitLab Docker Kubernetes Helm Flux CD Results ✅ Deploy time: from 2 hours to 10 minutes (12x) ✅ Errors: −90% ✅ Deploy frequency: from 1/week to 10+/day ✅ Rollback time: from 1 hour to 2 minutes

Read

Internet Radio on AzuraCast

Recovery and deployment of internet radio from scratch Client Community of independent artists with a mobile application and an audience of several dozen daily listeners Challenge After losing a rented VPS (along with the active configuration, website, and broadcast history), it was necessary to completely restore broadcasting for two internet radio stations. The only input data was a set of mp3 files. The task required deploying a new server, setting up an automated broadcasting platform, recreating a landing page with an embedded player, and ensuring stable 24/7 operation.

Read

Migration and DevOps for a CS-Cart Marketplace

Infrastructure and Deployment of a Multi-Vendor CS-Cart Marketplace Client Multi-vendor marketplace Challenge Revive a legacy marketplace project and migrate it from standard shared hosting to a modern cloud environment. Ensure high availability, build CI/CD infrastructure, set up test environments, integrate third-party delivery services, and automate routine operations with AI. At the final stage, prepare an on-premise server (NAS) on the client side for cold repository storage. Solution 1. Infrastructure in Yandex Cloud (Terraform) Built IaC configuration (Terraform) to provision Yandex Cloud infrastructure (virtual networks, VMs). Configured separate virtual machines for production and development environments. Attached and partitioned additional disks, deployed automated backups for databases and files to Yandex Object Storage (S3). Used Yandex Cloud Postbox for reliable service email delivery (domain, SPF, DKIM configured). 2. Containerization and Microservices Migrated the marketplace from classic hosting to a Docker-based VPS environment. Built a custom Dockerfile for CS-Cart and a docker-compose stack with a microservice-oriented web server architecture. Established full-featured dev and prod environments. Configured deployment pipelines via a local Git server (Forgejo). 3. Marketplace Platform (CS-Cart) Upgraded the CS-Cart core and optimized the database by removing obsolete, unused plugins. Deployed a clean platform installation with a new commercial theme. Configured multi-vendor business logic: role separation and seller dashboards. Integrated and configured modules: dynamic product loading, AI capabilities, smart live search, and SEO module. Integrated shipping (CDEK) and payment systems. Launched automated data exchange with 1C accounting systems. 4. Automation and n8n Deployed n8n process orchestration. Integrated ChatGPT API to automate repetitive operations such as content generation and normalization. Configured workers and triggers for store business workflows. 5. Security and Cloudflare Configured Cloudflare Proxy DNS. Implemented Cloudflare Turnstile to reduce spam during sign-up and checkout flows. Added bot protection: disabled direct registrations, configured strict caching, and filtered low-quality traffic. Installed SSL certificates and configured proper redirects. 6. On-Premise Infrastructure and Documentation Set up a dedicated subdomain with a wiki engine for technical and user documentation. Implemented local infrastructure in the client’s office based on Synology NAS: Docker, S3 buckets, and backups. Deployed a local Git server (Forgejo) with a CI/CD runner for version control and long-term preservation of project assets. Collected a complete backup of Terraform states, passwords, codebases, and configurations. 7. Project Archival and Repositories Migrated the production site to backup hosting for archival mode and disabled active synchronizations. Procured and configured Synology NAS in the office: external access, Docker, S3 buckets. Deployed Forgejo with a runner under the git.* domain. Repositories included: website source code, theme and add-ons, documentation, n8n workflows, Telegram bots, DB backups, terraform-yandex, terraform-cloudflare, and password vault. Technologies Docker Terraform Yandex Cloud Nginx Forgejo Linux Results ✅ Infrastructure: shared hosting → VPS in Yandex Cloud via Terraform, fully containerized with Docker ✅ Platform: clean CS-Cart Multi-Vendor installation replaced a heavily modified legacy core ✅ Theme and modules: commercial theme, live search, UX/UI improvements, SMS authentication, AJAX catalog ✅ Integrations: CDEK, Boxberry, Russian Post, Faster Payments System (SBP), and dual-environment 1C exchange ✅ Documentation: wiki on docs.* subdomain with a complete project structure ✅ Automation: n8n + ChatGPT API for routine operations ✅ Backups: Yandex Object Storage + mirrored copy on office Synology NAS ✅ Git: self-hosted Forgejo on NAS with runner and full project repositories ✅ Archival handover: project fully transferred to the client’s on-premise Synology NAS, including a private Forgejo Git server.

Read

Project Management: recovery and launch of a multi-vendor marketplace

Development Management for a Multi-Vendor E-commerce Project Client Startup — a multi-vendor marketplace. Challenge The marketplace had been deeply frozen for about two years. Before the freeze, the team attempted to build a custom template, but it remained unfinished. My initial goal as Tech Lead / Project Manager was to recover the project, hire and coordinate a distributed team (1C developers, web developers, SEO contractors, marketers), and bring the product to a working state end-to-end: from storefront to checkout.

Read

Prometheus + Grafana Monitoring Stack

Observability Stack for Microservices Architecture Client Early-stage startup Challenge After migrating to a microservices architecture (15+ services), the team had no centralized monitoring in place. Issues were only discovered through user complaints — typically 30+ minutes after they occurred. A full observability stack was needed to detect and diagnose problems proactively. Solution 1. Monitoring Architecture Prometheus for metrics collection Grafana for visualization Loki for centralized log aggregation Jaeger for distributed tracing Alertmanager for notifications 2. Metrics Collection Automatic service discovery in Kubernetes Application-level custom metrics System metrics via node-exporter Database metrics via postgres-exporter and redis-exporter 3. Grafana Dashboards Per-service dashboards for each microservice Unified infrastructure overview dashboard SLA/SLO tracking metrics Business metrics (RPS, conversion rate) 4. Centralized Logging (Loki) Log aggregation across all services Full-text log search via Grafana Log-to-metric correlation 5. Distributed Tracing (Jaeger) HTTP request tracing across services Call chain visualization Bottleneck identification Per-service latency analysis 6. Alerting Alerts delivered to Slack / PagerDuty / custom webhooks Critical issue escalation On-call rotation support Automatic incident creation Technologies Prometheus Grafana Kubernetes Docker Helm Linux Results ✅ MTTD: reduced from 30 minutes to under 1 minute ✅ MTTR: recovery time reduced by 60% ✅ Alerts: proactive notifications before users are impacted ✅ Visibility: full observability across all services ✅ Capacity planning: data-driven resource forecasting

Read

Self-Hosted Cloud Storage with Nextcloud

Self-Hosted Corporate Cloud Client Mid-sized business with strict data privacy and data residency requirements Challenge The company relied on third-party cloud services to store and share work files, creating data leakage risks and dependency on external providers. They needed a self-hosted solution with in-browser document editing, deleted file recovery, revision history, and additional collaboration tools — calendar, notes, and email — all under their own control. Solution 1. Nextcloud AIO Deployment Nextcloud All-in-One — official Docker image with the full stack out of the box PostgreSQL for application data storage Redis for caching and background job queues Nginx as a reverse proxy with automatic SSL/TLS 2. In-Browser Document Editing Nextcloud Office (Collabora Online) — built-in office suite Support for .docx, .xlsx, .pptx and ODF formats Real-time collaborative editing No local software installation required 3. File Management Deleted files trash bin with configurable retention period File version history — roll back to any previous revision Granular access control: folders, share links, passwords Mobile and desktop sync clients 4. Collaboration Tools Calendar (CalDAV) — team scheduling and events Notes — personal and team notes with Markdown support Mail — built-in web client for corporate email (IMAP/SMTP) 5. Backup Borg Backup — incremental backup built into Nextcloud AIO Data deduplication and compression Scheduled automatic runs Remote storage for backups (S3-compatible / SFTP) Technologies Nextcloud Docker PostgreSQL Nginx Linux Results ✅ Data ownership: files stay on company servers, no third-party providers ✅ In-browser editing: office documents open and edit directly in the browser ✅ Safety net: version history and trash bin protect against accidental data loss ✅ Unified platform: single solution for files, calendar, notes, and email ✅ Automated backups: incremental backup via Borg Backup on a set schedule

Read

Self-Hosted Corporate Messenger (Matrix)

Self-Hosted Secure Messenger for Enterprise Communications Client Mid-size business with strict data privacy and security requirements Challenge The company required full control over its internal communications — no third-party servers, no data leakage risks. The solution had to support end-to-end encryption, voice and video calls, file sharing, and seamless integration with existing corporate infrastructure, all manageable by an internal team. Solution 1. Server Stack Matrix Synapse as the core messaging server PostgreSQL 16 for persistent data storage Caddy as reverse proxy with automatic SSL/TLS Docker Compose for service orchestration 2. Client Applications Element Web for browser access Element Desktop for Windows/macOS/Linux Element Mobile for iOS/Android Consistent interface across all platforms 3. Voice & Video Calls Coturn (TURN/STUN server) for NAT traversal Group video call support UDP ports 49160–49200 for media traffic Automatic configuration via environment variables 4. Administration Synapse Admin web UI for user and room management Usage statistics and monitoring Accessible on a dedicated port (8888) 5. Security End-to-end encrypted messages Automatic SSL/TLS certificates via Caddy Public registration disabled Optional federation with other Matrix servers Healthchecks on all services 6. Automation Single Bash script for full stack initialization Automatic Synapse config generation Automated admin user creation via expect Docker Compose with dependency ordering and healthchecks Technologies Docker PostgreSQL Caddy Bash Linux Element + Matrix + Synapse Results ✅ Data ownership: full control over messages and user data — no third-party servers ✅ Scale: 100+ concurrent users ✅ Features: text, voice, video, file sharing up to 1.5 GB, E2E encryption ✅ Speed: full deployment in 5 minutes with a single script ✅ Reliability: automatic SSL certificates, healthchecks, auto-restart

Read

Self-Hosted Project Management with Plane

Self-Hosted Project Management with Plane So Client A company seeking its own project management tool hosted on their office infrastructure Challenge The client wanted an open-source alternative to Jira/Linear for project and task management, deployed on their own server within an existing Coolify environment. Requirements included: installing the latest version of Plane CE, ensuring proper operation behind Traefik reverse proxy, extracting the database as a separate Coolify service for convenient archiving, and setting up regular backups for all data.

Read

WireGuard VPN Infrastructure

WireGuard VPN Infrastructure Challenge A distributed team of 15+ engineers working across multiple locations needed secure, low-latency access to internal corporate infrastructure — without relying on heavyweight legacy VPN appliances. The existing setup introduced significant overhead, was difficult to maintain, and caused frequent connectivity issues for remote employees. The goal was to replace it with a lightweight, self-hosted solution that is fast, auditable, and easy to scale. Solution 1. Technology Selection Protocol evaluation: OpenVPN, WireGuard, Outline WireGuard chosen for its speed, simplicity, and modern cryptography Docker for isolation and portability Ansible for automated provisioning 2. Infrastructure VPS hosted in a neutral jurisdiction (Netherlands) Docker Compose for service orchestration WireGuard running in a container Nginx for the web management panel Prometheus + Grafana for monitoring 3. Automation # Ansible playbook for deployment - name: Deploy WireGuard VPN hosts: vpn_servers roles: - docker - wireguard - monitoring - backup 4. Security Automatic key rotation Firewall rules (UFW) Fail2ban for brute-force protection ChaCha20-Poly1305 traffic encryption 5. Monitoring Bandwidth and throughput metrics Downtime alerts Connection logging Automatic restart on failure Technologies WireGuard Docker Ansible Prometheus Results ✅ Uptime: 99.8% over 6+ months of operation ✅ Speed: stable 100+ Mbps ✅ Access: OpenAI API, ChatGPT, GitHub Copilot, npm registry ✅ Deployment time: ~20 minutes per new server ✅ Users: 15+ developers with zero connectivity issues

Read