Blog
Practical notes on running servers and websites.
How to find out why a disk is full, what load average really says, when a certificate expires — articles for people who are responsible for servers, with commands that have been checked.
Articles
Online shop down at night: notice it before customers do
Why a home page that returns 200 is not enough for a shop: check cart and checkout for a keyword, watch certificates, use several locations, escalate.
AI root cause analysis for server outages: what works
How to use a language model for root cause analysis of a server outage: what evidence to give it, what to redact, a prompt structure and how to check it.
How to check installed Linux packages for CVEs
Check installed packages for known CVEs on Debian, Ubuntu and RHEL with debsecan, pro, OVAL and dnf updateinfo – and decide what to fix first.
Monitoring client servers as an agency or hosting provider
How agencies, MSPs and hosting providers monitor many client servers: one area per client, least privilege, safe agent rollout, alert routing, offboarding.
MySQL binary logs filling the disk: purge them safely
MySQL or MariaDB binary logs eating your disk? How to find them, purge them with PURGE BINARY LOGS instead of rm, and set a retention that fits.
Status pages for your clients: a setup guide for agencies
How agencies, hosting providers and MSPs set up status pages for clients: one page per client, a custom domain via CNAME, TLS, subscribers and maintenance.
Uptime monitoring without false alarms: why locations matter
Why a single check location causes false alarms, how majority decisions and keyword checks fix that, and how to verify an outage yourself with curl.
Alert fatigue: debounce, quiet hours and escalation
How to cut monitoring noise without missing outages: debouncing, recovery notices, quiet hours, escalation and alerting on consequences, not causes.
Cron job monitoring: catch silent failures with a heartbeat
Why cron jobs fail without anyone noticing, and how the heartbeat pattern with a single curl call tells you when a backup or script did not run.
Linux disk full: how to find what is using the space
Disk full on a Linux server? Find the cause with df, du and ncdu, free space held by deleted files, and clean up safely without breaking services.
Linux load average explained: what is too high?
What the three load average numbers mean, how to relate them to the number of CPU cores, and how to tell CPU load from I/O wait on a Linux server.
PHP-FPM: server reached pm.max_children – what to do
What the PHP-FPM warning 'server reached pm.max_children' means, how to calculate a safe value from memory, and how to check the pool status.
SSH hardening checklist for Linux servers
Harden SSH on Debian and Ubuntu step by step: keys instead of passwords, no root login, drop-in config, firewall, fail2ban, testing without lockout.
TLS certificate expiry: how to monitor it properly
Why certificates expire even with certbot, how to check the certificate your server really serves with openssl, and how to get warned in time.
Would you rather be told than look it up?
Lynx Heartbeat watches the things these articles are about — and tells you before they become an outage.