PHP errors
The PHP errors screen shows the fatal errors and warnings PHP raised on a WordPress site: what went wrong, in which plugin, theme or file, how often, and the change on the site it followed. A fatal error stops the page it happens on, so a visitor or an editor sees an error instead of the page; a warning lets the page carry on, but often means something does not work as it should.
Open it from PHP errors under Monitoring in the site menu. The errors come from the Scanfully plugin on the site, version 1.12 or newer, so this screen needs the site to be connected with the plugin.
How Scanfully catches errors
Section titled “How Scanfully catches errors”The Scanfully plugin catches the errors PHP raises while WordPress runs, on the front end, in the admin, in AJAX and REST requests, in scheduled tasks and in WP-CLI. Notices and deprecation messages are left out.
The plugin groups the same error into one entry, so an error that happens on every page view is one row with a count, not thousands of rows. Two errors are the same when they have the same level, file, function and message; numbers and other details that change from one occurrence to the next are left out of the message when they are compared.
New fatal errors are sent to Scanfully within seconds. Counts of errors Scanfully already knows are sent every hour.
Coverage
Section titled “Coverage”To catch errors from the very first moment WordPress loads, the plugin installs its small must-use plugin, scanfully-mu.php, in wp-content/mu-plugins. When the plugin cannot write to that folder (some hosts do not allow it), it catches errors from the moment the Scanfully plugin itself loads, and tries again each time someone opens the WordPress admin, so once the folder is writable the must-use plugin is installed. The screen then says that coverage is partial: a fatal error while another plugin loads before the Scanfully plugin is missed. The must-use plugin shows on the Plugins screen as Scanfully’s own.
Privacy
Section titled “Privacy”The plugin sends the error message, the file and line, the function, a short stack trace (at most 10 calls, without the values passed to them) and the kind of request with its path, without the query string. Paths are relative to the WordPress folder, so the server’s own folders are not sent.
The cards
Section titled “The cards”- Open fatal errors: the fatal errors that keep happening.
- Open warnings: the warnings that keep happening.
- Came back: the errors that came back in the last 24 hours after 4 quiet days.
An error is open while it keeps happening: new, ongoing or came back.
The error list
Section titled “The error list”The tabs above the list choose which errors you see: Open, only Fatal errors or only Warnings, Resolved and Ignored errors, or All. From the Plugins screen, the error count of a plugin opens the list with only that plugin’s errors.
Each row shows:
- the level (Fatal or Warning) and the status;
- the message, and where it came from: the plugin, theme, must-use plugin or drop-in by its name and version, WordPress for WordPress itself;
- a small graph of the last 7 days;
- how often it happened, in all and in the last 24 hours;
- when it was first and last seen;
- the change it followed, such as “right after Example Forms was updated to 2.4.1”.
Statuses
Section titled “Statuses”| Status | When |
|---|---|
| New | First seen in the last 24 hours. |
| Ongoing | Keeps happening after its first day. |
| Came back | Happened again after 4 days without it, in the last 24 hours. |
| Resolved | Not seen for 4 days. |
| Ignored on this site or Ignored on every site | You ignored it. |
| Only during an update | Seen only while WordPress was in maintenance mode (see below). |
Errors while WordPress is in maintenance mode, the few seconds an update replaces files, are expected. They are counted and shown, but an error seen only during an update is resolved, and becomes new when it first happens outside maintenance mode.
The change it followed
Section titled “The change it followed”For each error Scanfully looks in the activity log for the change just before it: an update, an install or activation, a theme switch, a file edited in the WordPress editor, or a must-use plugin added or changed, from an hour before the error appeared until a few minutes after it (the activity of an update can arrive shortly after the error it caused). Only a change to the error’s own plugin, theme or must-use plugin, or a WordPress update, is named:
- Right after: the change was at most 10 minutes before the error. That change usually caused it.
- After: the change was 10 minutes to an hour before the error.
A change to another plugin or theme is not named, even when it was the closest one.
Error details
Section titled “Error details”Click a row, or the arrow in front of it, to see the error in full:
- the whole message, the file and line, and the function it happened in;
- the stack trace: the calls that led to the error, the latest first;
- the request it happened in (front end, admin, AJAX, REST, scheduled task or WP-CLI), its path, and the PHP and WordPress versions;
- a graph of the last 7 days, per hour;
- the versions of the plugin or theme it was first and last seen in.
Ignore an error
Section titled “Ignore an error”Some warnings are harmless and come from a popular plugin you cannot change. Ignore them so they stop counting:
- Ignore on this site ignores the error on this site.
- Ignore on every site ignores the same error on every site of your account, also on sites where it appears later.
- Stop ignoring on this site counts the error again here. It also ends the every-site rule, so the error counts again wherever it appears next; the other sites keep their own ignore until you stop it there.
- Stop ignoring on every site counts the error again on every site.
An ignored error never sends a notification and is left out of the counts, the Site health issue and the digest. You can also ignore an error with the ignore_error tool of the Scanfully MCP server or the public API.
Notifications
Section titled “Notifications”A notification with the New fatal PHP error or Fatal PHP error came back event, in the Plugin health category, tells you when a fatal error is new on a site, or comes back after 4 days away. You hear about each error once: the same error happening again does not send another alert. When the error followed a change to its own plugin, theme or must-use plugin, or a WordPress update, the alert says so, for example “It appeared right after Example Forms was updated to 2.4.1.” (see The change it followed).
The Everything else notification every account has includes both events. Warnings do not send alerts; they show on this screen, and in the daily digest when a fatal error changed. See Notifications for how to choose who gets which alerts.
A fatal error that takes the whole site down is also caught by the uptime monitor, which names it as a WordPress critical error.
Elsewhere in Scanfully
Section titled “Elsewhere in Scanfully”- Site health shows a critical issue while fatal errors happened in the last 24 hours, with the plugins or themes they came from.
- Plugins shows the open fatal errors and warnings of each plugin next to it.
- The System tab of the activity log has a row each time an error is new or comes back, never one per occurrence.
- The daily digest lists the fatal errors that were new or came back that day, and how many errors are open.
How long errors are kept
Section titled “How long errors are kept”Hourly counts are kept for 90 days. An error not seen for 90 days is removed. A site keeps at most 1,000 errors; when it reaches that, the oldest resolved or ignored error makes room, and when every error is open, new ones are only counted.
