Moving one WordPress site is easy. Moving twenty in a weekend is a pipeline, and the failures that cost you are the quiet ones: writes landing on the old server, cron firing on two boxes, certificates blocked by ACME rate limits. Here is the fleet migration process I use, from manifest to per-site rollback.
Both cron and systemd timers can run a command at 2 a.m. The difference is what happens the night it doesn't run, and how long it takes anyone to notice. A comparison by failure family, with a decision rule, the same job written both ways, and how to debug each one when it goes quiet.
DirectAdmin gives each user one PHP-FPM pool, not each site. That single design fact explains most of the 502s on a busy shared box. Here is how to find the pool files, size pm.max_children from real memory numbers, set per-site limits that WordPress cannot override, and make the changes survive the next rewrite_confs.
High I/O wait is a symptom, not a diagnosis. It rises when nothing is wrong and stays flat when something is badly wrong, and it only sees one layer of the storage stack. This is how to read iowait and PSI correctly, then work through the five failure families that actually produce storage stalls: device saturation, memory pressure, synchronous writes, throttling, and latency that lives above the block layer entirely.
Migrations rarely fail loudly. The site stays up while mail, cron, certificate renewal and the write path break silently behind it. A post-migration checklist for the first 48 hours, ordered by how long each failure can hide rather than by subsystem.
A single Ansible run can take a blank VPS to a hardened, reproducible server. The hard part is not the tasks, it is the order. Here is the ordering that keeps you logged in, the failures that pass green and bite two weeks later, and how to make the second run as safe as the first.
Benchmarks pick the wrong winner. Here is how Nginx, Apache and Caddy actually differ in production: how each one fails under load, who owns certificate renewal, and where configuration lives when a developer needs to change it. Includes a decision procedure and the mistakes I see most.
Hand-built client servers fail quietly: you only find out they were never reproducible on the day you need to rebuild one. Here is how to automate server provisioning for new clients using cloud-init for first boot, Ansible for convergence, per-client secrets and a verification gate that refuses to hand over a half-built box.
Docker's default logging driver writes container output to a JSON file with no size limit and no rotation. This is a practical guide to Docker log rotation: what the defaults actually do, why your daemon.json change did nothing, why deleting the log file did not free any disk, and how to choose between json-file, local, journald and shipping logs off the box.
The dashboard is fine for editing a post. It is a terrible tool for anything touching a thousand rows, twelve plugins or two servers. Here are the WP-CLI commands that save hours on migrations, plugin conflicts, stalled cron and integrity checks, grouped by the job that eats the time, with the failure modes each one hides.