Introducing Matmul as a Service, backed by tinygrad

We pivoted from server monitoring to selling matrix multiplications over HTTP. Here's how we built a tinygrad runtime that runs compute one HTTP request at a time.

Engineering

MCP design is prompt engineering all over again

After building an MCP server for Simple Observability, we realized MCP design has a lot more in common with prompt engineering than traditional API design.

How we built a metric simulator

How we built a deterministic metric simulator at Simple Observability to replace hardcoded test data.

Linux kernel doesn't care about your disk health

Why disk health monitoring on Linux is a mess of legacy protocols, kernel permission traps, and version-specific workarounds.

Lightweight is a metric, not an adjective

How we define and measure the dynamic footprint of simob. A deep dive into /proc, memory attribution, and tracking resource usage in CI.

Go is portable, until it isn't

An exploration of Go's portability limits when integrating with system-specific C libraries like libsystemd and dealing with glibc vs. musl systems.

Why server monitoring matters?

Why monitoring your app isn’t enough. Discover how server monitoring helps you catch issues early, stay secure, and respond fast.

Blocking web scrapers: A realist’s approach

A realist’s guide to blocking web scrapers with layered defenses. Practical strategies from real-world experience to slow down automated data extraction.