Article

The WordPress maintenance checklist for agencies: daily, weekly, monthly

Michal

Michal

Author

6 min read
The WordPress maintenance checklist for agencies: daily, weekly, monthly

If you look after WordPress sites for other people, maintenance is the part of the job nobody sees until it goes wrong. A plugin update takes a checkout page down on a Saturday. A certificate lapses on the one site that is not behind Cloudflare. A backup you assumed was running has been failing quietly for six weeks. None of these are hard problems; they are forgotten problems.

This checklist is the one we use across a fleet of just over a hundred client sites. It is split by how often each item needs doing, and it marks which items a tool should do for you and which still need a person to look. If you only take one thing from it: attach the list to a schedule and to a tool that does the checking, or it will stop being a list and go back to being a vague worry.

Daily: the ten-minute pass

Most of the daily work should happen without you. The point of the daily pass is to look at what the automation found, not to go looking yourself.

Automated, every day

  • Uptime and SSL. Every site checked at least once a minute, with the certificate expiry date read on each check. A site that is "up" but serving an expired certificate is down for most visitors.
  • Backups ran and verified. Not "a backup job was scheduled", but "a backup finished, its checksum matches, and it is stored somewhere other than the web server".
  • Vulnerability feed. Installed plugin and theme versions matched against a public vulnerability database. A critical vulnerability in a plugin you have on twelve sites is a today problem, not a next-update-window problem.
  • PHP fatal errors. Collected from each site's error log and grouped, so a new fatal after yesterday's update stands out from the usual deprecation notices.
  • Login activity. Failed logins over a threshold and successful logins from a new country. Brute-force attempts are background noise; a successful login from somewhere new is not.

A person, ten minutes

  • Anything down, anything vulnerable, anything whose backup failed. Those three categories go first, every day, before anything else.
  • Approve the low-risk updates that queued overnight. If your tool scores risk (patch release, no database changes, no template changes), these should be a one-click batch.
  • Glance at the sites that stopped syncing. A site that has not reported in for a day usually has a broken plugin, a changed password, or a host that blocked your dashboard's IP.

Weekly: updates and housekeeping

  • Plugin and theme updates, applied in a maintenance window with a snapshot before and a check after. Security fixes first, major versions last and one site at a time. There is a separate post on doing this safely at scale.
  • WordPress core minor releases. Minor releases are security and bug fixes and are safe to apply the week they land. Major releases wait for the monthly review.
  • Cron health. WP-Cron only runs when someone visits the site. Low-traffic client sites end up with overdue jobs, failed scheduled posts and WooCommerce Action Scheduler backlogs. Check for jobs overdue by more than twice their interval and for orphaned hooks left behind by deactivated plugins.
  • 404s and broken links. A weekly look at the top 404s catches a renamed page before the client's customers do, and catches bots probing for files that should not exist.
  • Database bloat. Expired transients, post revisions in the thousands, orphaned post meta and WooCommerce sessions. A cleanup takes seconds and keeps the autoload size under control, which is the single most common cause of a slow admin.
  • Error log review. Anything that appeared this week and was not there last week.

Monthly: the review

  • PHP version. Each PHP branch gets two years of active support and then security fixes only. Know which sites are on a branch that is end-of-life or about to be, and plan the upgrade with the host while it is still routine.
  • Abandoned plugins. Plugins closed or removed from the WordPress.org directory still run on your sites; they just never get another update. Replace them while the replacement is a choice and not an emergency.
  • Administrator accounts. Who has an admin account on each site, and should they still? Former staff, old agencies and "temporary" developer accounts accumulate.
  • Performance. Core Web Vitals per site, on mobile. One number per site per month is enough to see which direction things are heading.
  • Domain expiry. Domains are renewed by clients, at registrars you do not control, with cards that expire. A WHOIS check once a month, with a 30-day warning, has saved more than one site.
  • The client report. Updates applied, uptime, backups taken, issues fixed. Clients do not see maintenance unless you show it to them, and the report is what justifies the retainer.

Quarterly: the things that only matter when they matter

  • Restore a backup. Not to production: to a staging site or a local install. A backup you have never restored is a hope, not a backup.
  • Licences. Premium plugin and theme licences that lapse stop receiving updates, and the first you hear of it is a vulnerability you cannot patch.
  • Unused plugins and themes. Deactivated plugins are still code on the server and still get scanned, still get exploited. Delete them.
  • Hosting review. Disk usage, PHP memory limits, whether the host still supports the PHP version you need next year.
  • Maintenance windows and alert routing. Who gets woken up, for which sites, and whether that is still right.

Turning the list into a routine

The list above is long, and that is the problem with most maintenance checklists: nobody runs a fifty-item checklist by hand across eighty sites. Three things make it workable.

  1. One place that already knows the answers. The daily automated items should land in a single queue, ranked by severity, so the ten-minute pass is reading, not hunting. If you are opening wp-admin on each site to see what needs updating, the tool is doing the easy part and leaving you the hard part.
  2. Automate what has a safe default. Low-risk updates with a snapshot and rollback, daily backups with verification, uptime checks with confirmation from a second location. These do not need judgement, so they should not wait for yours.
  3. Report what you did. The monthly report is the only time most clients think about maintenance. Make it automatic too, or start from our free WordPress maintenance report template.

This is the shape WP Admin Online is built around. The free plan covers the monitoring half of the list for unlimited sites: the attention queue, vulnerability detection, SSL and domain expiry, PHP end-of-life, error collection and one-click login. The paid plans add the automation: safe updates with rollback, uptime monitoring, backups and client reports. Either way, the checklist stops being a document and becomes a list of things that are already done.

Share this article

Help others discover this content