Building a GitOps Homelab Platform Pipeline
Manual pod updates became a GitOps platform pipeline where CI builds images, Kustomize separates dev and prod, ArgoCD reconciles drift, and short-SHA tags trace rollbacks.
- #kubernetes
- #platform
Manual pod updates became a GitOps platform pipeline where CI builds images, Kustomize separates dev and prod, ArgoCD reconciles drift, and short-SHA tags trace rollbacks.
Transitioning to Agentic Infrastructure via the Model Context Protocol unified gateway, providing specialized tools for automated RCA across telemetry, pods, network flows, and gateway capability state.
Transitioning from imperative Makefile scripting to a declarative OpenTofu orchestration layer for managing a growing Kubernetes observability stack through Infrastructure as Code.
Rebuilding a Linktree-style Vue project as a Go static site generator to reduce npm maintenance, simplify builds, and learn Go through a small real-world publishing tool.
Building an MCP server with Go and SQLite to give AI agents scoped, local-first memory, reduce repeated context setup, and keep project decisions available in CLI sessions.
Fixing Grafana dashboard provisioning by moving from brittle embedded JSON in values.yaml to standalone files, resolving 'invalid character' bugs once and for all.
Transitioning from fragmented multi-module scripts to a library-first platform using a unified Go workspace.
Mastering SRE principles by transitioning from ad-hoc data collection to a standardized OpenTelemetry ecosystem using Prometheus, Grafana Tempo, and MinIO for persistence.
Executing a risk-based, four-stage migration of the observability stack from Docker to Kubernetes to ensure data integrity and system stability.
Migrating from EOL Promtail to Grafana Alloy via a Strangler Fig pattern, bridging Kubernetes and Docker logs to ensure zero visibility loss during the platform transition.