How-tos and explanations for Linux, servers, databases and the cloud. Every how-to was run on a fresh machine first, and every piece says what it does not tell you.
measured our run on real machinesmodel the simulatorexplained a concept, with one simulated exampletested every command run by us on a fresh machine
gpt-oss-20b on one rented L4, measured: what a million output tokens cost at 1 to 64 requests at once, and how busy the GPU must be to beat five per-token APIs.
1,008output tokens per secondone L4, gpt-oss-20b, 64 requests at once
A request that waits for 100 parallel calls is as slow as the slowest of them: the arithmetic of Dean and Barroso's paper, and the same effect in a design you can run.
A $12-a-month VM with Nginx, Node.js and Postgres under virtual users: about 800 users at one second of think time, and a CPU throttle the guest does not show.
One PostgreSQL on a 2-vCPU VM, measured: about 6,000 simple indexed reads per second at 100 % CPU, where p99 starts to climb, and how to estimate your own number.
We installed Docker Engine on a fresh Ubuntu 24.04 and kept each output: a first container, “permission denied” on the socket, the uninstall, and what stayed behind.
We installed PostgreSQL on a fresh Ubuntu 24.04 and kept what every command printed, from the first login and “Peer authentication failed” to removing it all again.