Year 2019 — A Recap of Our Cloud and Kubernetes Journey

The end of the year is a time to take stock. From the first experimental Kubernetes 1.5 cluster in January 2017, we’ve grown to three production clusters, a multi-cloud strategy, and fully automated CI/CD. A three-year recap.

Where We Were in 2017¶

On-premise servers, manual deployments, Nagios monitoring, Jenkins freestyle jobs. Kubernetes as an experiment. Docker as a novelty. Cloud as a distant dream for banking clients.

Where We Are Today¶

• 3 production Kubernetes clusters: on-premise (Prague), GKE (Frankfurt), AKS (West Europe)
• 50+ microservices running in containers
• GitOps with ArgoCD for continuous delivery
• Prometheus + Thanos for multi-cluster observability
• HashiCorp Vault for secrets management
• Apache Kafka for event-driven architecture
• Terraform for all infrastructure

Key Lessons¶

1. Invest in people. Kubernetes and cloud technologies change fast. Training, conferences, certifications — the most important investment.

2. Automate everything. Manual steps are errors waiting for an opportunity. CI/CD, IaC, GitOps — automation is the foundation of reliability.

3. Observability first. You can’t fix what you can’t see. Monitoring, logging, tracing must be part of every service from day one.

4. Security is everyone’s job. DevSecOps, shift left, compliance as code. Security is not an afterthought.

5. Start small, iterate. Don’t do a big bang migration. Start with one service, learn, iterate.

Outlook for 2020¶

Service mesh in production (Istio or Linkerd). Serverless for event-driven workloads. FinOps maturity. Multi-cluster federation. And most importantly: helping more clients on their cloud journey.

Cloud Transformation Is a Marathon, Not a Sprint¶

Three years of work, hundreds of hours of learning, dozens of mistakes and lessons. But the result is worth it — our platform is more robust, flexible, and scalable than ever before. And this is just the beginning.