I managed up to seven servers hosting applications and services for MetaZenIT.
Up to 7 servers
Peak size of the infrastructure I managed at MetaZenIT.
How a change reached production
Simplified architecture · GitHub Actions ran integration jobs. Kargo promoted changes and Argo CD reconciled the cluster with the state declared in Git.
Engineering decision
Finding a difference between Git and the cluster
The manifest recorded what should be running. During storage or connectivity incidents I compared that configuration with operational metrics in Prometheus and Grafana.
Read the full technical study
My responsibility
I managed Linux, networking, DNS and TLS. I packaged applications with Docker, deployed them with Helm and operated Kubernetes and K3s clusters through Rancher.
From repository to production
GitHub Actions handled continuous integration. Kargo managed promotions and Argo CD reconciled the manifests with the cluster. Ansible automated server tasks.
Operations and lessons
I investigated storage and connectivity incidents using Prometheus and Grafana. Keeping desired state in Git gave me a reference when comparing a failing service with its intended configuration.