Skip to content
Menu

Monoliths to highly available Kubernetes

Role: Team lead and senior engineer2022 – 2023RusBITech-Astra

Internal details are anonymised under NDA. Diagrams are redrawn and the numbers rounded.

  1. Monolithic service
  2. Containerised components
  3. HA Kubernetes
Redrawn, simplified flow

problem

Mission-critical internal services — source control, messaging and the corporate single sign-on — ran as monoliths. Each was a single point of failure for the engineers who depended on it.

what I did

  • Corporate SSO. I re-architected the Keycloak SSO from a monolith into a highly available, cloud-native service.
  • SCM and messaging. As team lead, I architected and ran the migration of the source control and messaging services to highly available, scalable services on Kubernetes.
  • Automation underneath. I refactored 20+ legacy Ansible roles into modular collections and built CI/CD for fast, predictable, zero-downtime deployments.
  • Observability. I set up monitoring and alerting with Prometheus, Grafana and Alertmanager.

I led and grew a unit of 8 engineers through this work.

Outcome

  • 8engineers in the unit I led
  • 20+legacy Ansible roles refactored into collections

Stack

KubernetesDockerKeycloakTerraformAnsiblePuppetPrometheusGrafana