JUN 2022 — PRESENT · Ljubljana, Slovenia
DevOps Engineer
A1 Slovenija, d.d.
Started with deploying applications and services on microservice principles, then went deep into the tooling that makes it possible: Git, Linux, Podman & Docker, Kaniko, Buildah and Source-to-Image. From there — Kubernetes manifests and deploying images to the OpenShift cluster in many different ways, including BuildConfigs with Dockerfile and S2I. Developing pipelines in Kubernetes-native Tekton taught me where persistent volumes, mounted secrets, config maps and the different persistent volume claim types really earn their place.
- helm · iac · gitops Learned Helm to parameterize CI/CD pipelines into reusable templates deployable across environments. Recognized the power of Infrastructure as Code and pushed the idea that every project and application should be of declarative state — maximizing ArgoCD and levelling up from DevOps to GitOps.
- tekton pipelines Wrote dozens of reusable Tasks and ClusterTasks — and more pipelines than I can count. Provisioned and integrated systems like our own SonarQube server into Tekton Tasks; currently researching HashiCorp Vault integration into CI/CD.
- feb 2024 · teaching Red Hat OpenShift Administration I — Managing Containers and Kubernetes. 3-month internal workshop.
- ansible · standardization Used Ansible to automate, provision and reuse templates as much as possible — leaving the human-error factor behind and learning the real importance of standardization.
- database platform migrationnew Planned and executed a mass migration of EDB PostgreSQL clusters on Kubernetes/OpenShift to Percona (Percona Operator for PostgreSQL). Target clusters were templated as Helm charts in line with our IaC standards, so every migrated database stays fully declarative and reconciled by ArgoCD. Migration steps were scripted and repeatable rather than performed by hand, allowing clusters to move between operators in a controlled, standardized way.
- current · platform engineering Having been on the developer's side, I know the pain of deploying to OpenShift. I'm actively building Byte Garden, our internal developer platform: self-service through dynamic forms rendered from Helm chart templates, one-pass provisioning of GitLab, Quay, Vault and ArgoCD, and the filled chart pushed to Git for ArgoCD to apply. Goal: almost no manual work — developers who feel self-sufficient and focus purely on writing code.