Occasional write-ups on networking, Linux and infrastructure.
ICMP echo is convenient but it is not the number your application sees. Once you add TLS handshakes, TCP slow start and a couple of middleboxes, the picture changes. A short note on what to measure, and why the first sample is almost always wrong.
For most small deployments the choice is narrower than the documentation suggests. A look at
what fq and fq_codel do differently under load, and why buffer bloat
shows up as latency long before it shows up as packet loss.
Unattended upgrades, key-only SSH, a firewall with four rules and a backup you have actually restored once. Notes on staying uninteresting to whoever is scanning port 22 this week.