Experimentation space
What I explore
Building to understand
No client constraints here. A space to test technologies, validate hypotheses, and go deeper than daily work allows.
5
experiments
2
ideas in progress
5
domains
What I work with
My tech stack
Technologies I use regularly, organized by domain.
Frontend
Backend
DevOps & Infra
Observability
Tools
Experiments
What I built to learn
Self-hosted monitoring
Full-stack monitoring with Grafana, Loki, Promtail and Prometheus on my Proxmox infrastructure. Uptime Kuma monitors service availability and sends uptime alerts to a dedicated Discord. Alertmanager handles system alerts like storage threshold exceeded and routes them to Discord as well.
What I learned
- Setting up automatic Docker service scraping with Prometheus.
- Building custom Grafana dashboards per service.
- Centralizing application logs with Loki and Promtail.
- Configuring Alertmanager with system alert routing to a dedicated Discord channel.
- Deploying Uptime Kuma with uptime notifications to a separate Discord.
Self-hosted CI/CD runner
Self-hosted GitHub Actions runner for my project deployments. CI/CD logs are sent directly into my observability stack for unified tracking.
What I learned
- Setting up and hardening a self-hosted GitHub Actions runner on Proxmox.
- Building zero-downtime Docker deployment pipelines.
- Integrating CI/CD logs into the Loki stack for centralized tracking.
Go CLI tooling
CLI tools in Go to automate repetitive tasks: project scaffolding, deployment, secrets management. Cobra for command structure, Viper for configuration.
Lightweight API Gateway
Functional Go API gateway with JWT authentication, configurable rate limiting and all security layers in place. Logging middleware, dynamic routing and route protection are implemented and operational.
What I learned
- Implementing JWT authentication with token validation and management.
- Setting up configurable per-route rate limiting backed by Redis.
- Building a structured logging middleware in Go.
- Dynamic routing to backend services with declarative configuration.
Nuxt SSR + Edge
Nuxt deployed on Cloudflare Workers (Edge) to minimize latency. Edge constraints are significant: no filesystem access, isolated workers and cold start overhead. Real technical challenges to work through.
On my mind
What I'm exploring next
Ideas waiting for free time. Some will become experiments, others might stay on hold.
Custom auth service
Building a full auth service from scratch: OIDC, rotating refresh tokens, device management and audit log.
Go ETL pipeline
Extract metrics from multiple heterogeneous services, normalize them and aggregate into a time-series database like InfluxDB. The goal is a unified monitoring view across different data sources.
My philosophy
Build to understand
The Lab is an unconstrained space. Every project, even abandoned halfway, is a chance to understand what tutorials never show.
Experiment
Test ideas without production pressure or delivery constraints.
Learn
Understand deeply, not just make something run.
Capitalize
Document what worked, what broke, and why.
Want to chat?
These explorations shape the way I think
Whether it's an architecture question, a technical challenge, or a tech I'm exploring - I'm always up for a conversation.