I reproduce configurations standard to financial-cloud operations, measuring downtime with probes, and analyze and systematize knowledge through an ontology lens — turning what I've experienced into what I understand.
| Lab | Stack | What it demonstrates |
|---|---|---|
| web-was web / WAS three-tier | NGINX · Tomcat · MySQL · Redis | Session experiments · fault injection · root-causing 502 vs 504 |
| k8s-oss Kubernetes zero-downtime | kind · Deployment · rolling update | Single-instance patch vs rolling update (maxUnavailable: 0) downtime comparison |
| observability Observability · SLA | Prometheus · Grafana · Python | Alert pending → firing lifecycle · availability ↔ allowed-downtime SLA toolkit |
Uses only public standards and synthetic data.
Ontology-based PKM · 4,300+ notes
A personal knowledge OS built on Obsidian, integrating ontology, PARA and AI. Notes are structured with a four-layer meta-model and ontology design, and workflows are automated through Claude Code and MCP.
Zero-downtime is a product of architecture — reproduced and measured
web/WAS three-tier, Kubernetes, and SLA/monitoring — a learning lab reproducing and measuring configurations standard to financial-cloud operations. Downtime for single-instance patching vs. k8s rolling updates was measured with probes. Uses only public standards and synthetic data.