Temper Cloud

Self-hosted infrastructure is often fragile: services are hard to expose securely, storage redundancy is an afterthought, observability is scattered, and pipelines that touch thousands of files become painfully slow. Most home-lab setups lack the Zero Trust security, monitoring depth, and throughput engineering that production environments demand.

Temper Cloud treats self-hosted infrastructure with production-grade rigour. RAID-backed distributed storage ensures data durability. Cloudflare Tunnel and mTLS-style private access eliminate exposed ports. A stateless API gateway (Caddy) routes traffic cleanly. Observability tooling (Netdata, Dozzle, cAdvisor, FastAPI status services) provides deep visibility. Metadata and embedding pipelines are engineered for concurrency, processing 10K+ files in under 4 minutes.

I architected a Docker-based distributed storage system with RAID-backed volumes around 3.9TB. I designed isolated service boundaries for each self-hosted component and built a Zero Trust access layer using Cloudflare Tunnel and mTLS-style private access patterns. I deployed a full observability stack with Netdata, Dozzle, and cAdvisor, plus custom FastAPI health/status services. The Caddy API gateway provides clean routing across all internal services. I also built concurrent metadata and embedding pipelines that process 10K+ files in under 4 minutes.

Expand storage capacity and redundancy. Add automated backup verification. Build a lightweight dashboard for service health and pipeline status. Document the Zero Trust access patterns as a reusable template for other self-hosted environments.