RSSAmplifier

Blog

Technical Partner for SMEs, Startups & Product Teams | Dr.-Ing. | Georg Schwarz

Founder of diffcraft, technical partner for SMEs, startups, and product teams, full-stack software engineer, and Doctor of Engineering (Dr.-Ing.), based in Nuremberg, Germany.

georg-schwarz.comRSS feed ↗16 posts

Latest posts

From Kubernetes Dev Setup to Production: What Actually Changes

How we moved a Kubernetes-hosted product from a development setup to a platform with controlled delivery, policy checks, observability, and tested recovery.

Scraping pricing data with AI: 5 Engineering Lessons

Lessons from building a pricing intelligence prototype that extracts semi-structured pricing data from live websites and turns it into decision-ready comparisons.

Serving Private S3 Objects: Backend Proxy vs Gateway Auth vs Presigned URLs

Compare three proven patterns for serving private files from S3-compatible storage and learn when to use backend proxying, gateway auth, or presigned URLs.

Zero-Downtime Ingress Controller Migration in Kubernetes

A step-by-step blueprint for migrating away from ingress-nginx with zero downtime. Uses parallel controller deployment and DNS-based traffic switching, applicable to any target controller.

OpenTelemetry Auto-Instrumentation in Node.js (with NestJS and NX)

How does OpenTelemetry auto-instrumentation actually work? A look behind the curtain at monkey-patching, module loading order, and why bundlers can break everything.

Exposing Applications from Private Networks Using Cloudflare Tunnel

Three deployment approaches for publishing applications from private networks via Cloudflare Tunnel: Docker Compose, Kubernetes with direct service routing, and Kubernetes with ingress controller integration.

Managing Multi-Chart Kubernetes Deployments

Learn how to deploy multiple Helm charts together using umbrella charts, custom scripts, Helmfile, and GitOps tools. Understand when each approach makes sense.

Monitoring DNS Issues in Kubernetes with Node Problem Detector

When random nodes lose DNS connectivity in a Kubernetes cluster, detection becomes critical. Here's how I deployed Node Problem Detector with custom DNS checks to catch these issues before they impact production.

Using Existing Helm Charts: Copy, Install, or Dependencies?

Learn the three approaches to using existing Helm charts in your Kubernetes deployments, understand their trade-offs, and discover which method works best for your use case.

Spegel vs Harbor: Which Pull-Through Cache Should You Choose?

A practical comparison of Spegel and Harbor for Kubernetes image caching. Learn how to evaluate these solutions based on operational complexity, architecture, and real-world deployment experience.

Personal Milestone: I defended my PhD Thesis about Microservices

After years of research into how microservices communicate and integrate, I successfully defended my dissertation "A Theory of Microservice Integration" at FAU Erlangen-Nürnberg.

Documenting Microservice Integration with MSAdoc

Documentation in microservice projects quickly becomes outdated. MSAdoc is an open source tool that enables decentralized documentation close to the code while aggregating it centrally for better discoverability.

Ensuring Syntactic Interoperability Using Consumer-Driven Contract Testing

Consumer-driven contract testing promises to catch integration issues early. Our research reveals how it prevents breaking changes and what teams need to know to implement it successfully.

Balancing Technology Heterogeneity in Microservice Architectures

One of microservices' biggest promises is technology freedom. But when does diversity become chaos? Towards finding a sweet spot.

A Taxonomy of Microservice Integration Techniques

Microservice integration is more than just technical challenges. Our latest research reveals different kinds of integration that include organizational and social aspects that are often overlooked but critical to success.

What Microservices Can Learn From Enterprise Data Integration

Microservices architecture has a blind spot: many proven data integration strategies from enterprise systems remain unused. Here's what your team is missing.