Build reliable k3s clusters by treating infrastructure creation and configuration as a deliberate contract.
Terraform to Ansible: Engineering the k3s Provisioning Handoff
Razvan Mares··8 min read
A k3s cluster can be running within minutes. A *repeatable* k3s cluster—one that survives rebuilds, node replacement, and the inevitable “what changed?” investigation—requires more than a short instal...
linuxproxmoxk3s+1
The infrastructure habits that turned side projects into systems I could trust.
DevOps Use Cases That Made My Personal Projects Easier to Run
Razvan Mares··8 min read
Personal projects have a way of exposing infrastructure problems at inconvenient times. A demo works perfectly on a laptop, then fails after a restart. A database migration succeeds locally but breaks...
PostgreSQLGitHub Actionsci-cd+3
Follow an asynchronous operation through the compiler, runtime, scheduler, and diagnostic pipeline.
Inside .NET 11 Runtime-Native Async: From Await to Continuation
Razvan Mares··8 min read
Async code is often taught as a source-level convenience: write `await`, avoid blocking a thread, and let the framework handle the rest. That model is useful, but it stops just before the interesting ...
.NETcsharprest-apis+1
A practical guide to making MCP servers discoverable, trustworthy, and maintainable.
From README to Registry: Publishing MCP Servers Agents Can Find
Razvan Mares··9 min read
The official MCP server registry changes a fundamental assumption about agent tooling: a GitHub repository is no longer enough.
For early Model Context Protocol (MCP) projects, discovery was informal...
GitHub ActionsnodejsTypeScript+2
Turn a C# MCP server into a clear, constrained interface that AI clients can use reliably.
Designing Discoverable MCP Tools in C#: Schemas, Boundaries, and Intent
Razvan Mares··9 min read
An MCP server is not merely an API with a new transport. It is a translation layer between probabilistic language models and deterministic software. That distinction changes how a C# developer should ...
.NETcsharpAI/ML+1
Create a paved runway from local code to dependable production without building an enterprise platform.
Platform Engineering as a Flight Deck for Solo Developers
Razvan Mares··9 min read
A solo developer does not need an internal platform team. They need the *outcomes* a good platform team creates: a safe path to production, repeatable environments, visible operational signals, and fe...