Caching
A site can keep two caches, both set in the Caching card on its Settings tab. The page cache stores finished pages in nginx, so most visitors never reach PHP. It is on by default. The object cache keeps WordPress’s database answers in Redis, which speeds up what the page cache skips, such as wp-admin and logged-in visitors. It is off by default.

Turn on the page cache
Section titled “Turn on the page cache”Under Page cache, choose On. Off turns it off for this site. Follow the default takes the switch from Settings → Sites, where it is on unless you change it. The change applies at once, without a restart, and empties the cache.
Only WordPress’s own pages are stored, for logged-out visitors, when they answer 200, 301 or 404. Files
on disk, scripts that do not load WordPress and pages a cache plugin serves before WordPress loads go
straight through. A page that says no-cache, sets a cookie or defines DONOTCACHEPAGE is not stored,
which keeps out WooCommerce’s cart, checkout and account pages.
When PHP or the database fails, visitors get the last stored copy of a page, if it has one. The panel’s own checks go past the cache, so they still see the failure.
A WooCommerce store on plain Geolocate shows prices by country at one address, so its pages are
never stored. Such a site goes to Off when it moves to this setup, unless it chose otherwise, and
the card says why. Geolocate (with page caching support) works with the cache, because each country
gets its own ?v= copy.
Clear the cache
Section titled “Clear the cache”Choose Clear cache on the site’s Overview tab. It empties the page cache, and Redis when it runs. Clear cache on WordPress’s admin bar, for administrators, empties the page cache only.
The cache also clears by itself:
- In WordPress, or under WP-CLI, when a published post or page changes, or a block theme’s templates, styles or navigation. So do approved comments, terms, menus, widgets, the Customizer, a theme switch, plugin activation, updates, WooCommerce stock and settings an administrator saves.
- When WP Rocket, W3 Total Cache, WP Super Cache, LiteSpeed Cache, Autoptimize or Elementor clears its own cache.
- After a panel job, or a change on the Files or Database tab, that can change pages. A WP-CLI command that only reads, or a GET through the REST tool, does not clear it.
- When a rule, the keep time or the page cache’s switch changes.
- Within 10 seconds of a change to the site’s root
.htaccess. - At every start of the container.
The panel cannot see edits in Adminer, FTP and SFTP changes to theme PHP files, or what runs in the server’s terminal. Use Clear cache after those.
Turn on the object cache
Section titled “Turn on the object cache”Switch on Object cache (Redis). A job starts Redis in the site’s container, installs and activates the Redis Object Cache plugin from wordpress.org, and turns on its drop-in. If the plugin cannot reach Redis, the job fails and Redis stops again. Switching it off stops Redis and deactivates the plugin, which stays installed.
Redis has no network port, so only the site’s own PHP and WP-CLI reach it. If it is down, the site runs without it. While another plugin’s object cache drop-in is installed, the switch stays off and names that plugin.
Skip the cache for a page
Section titled “Skip the cache for a page”Open Rules. It shows while the page cache is on, as do Keep pages for and Check a URL.

- Under This site’s rules, set What to Skip the cache.
- Set Matches, then fill in Value.
- Choose Add rule.
| Matches | Value |
|---|---|
| URL path | /shop/ is that page. /shop/* also takes what is under it, and /shop* whatever starts so. Case does not count, and any language’s letters work. |
| Cookie | The start of a cookie’s name. |
| Query parameter | A parameter’s name. |
| User agent | Text anywhere in the user agent, in its exact case. |
| Post type | A post type, such as product. WordPress applies this one, at the end of the page. |
A query parameter the cache does not know skips it anyway, so ?s= and ?add-to-cart= need no
rule. A visitor with a cookie it does not know can still get a stored page. So a login or membership
plugin with a cookie of its own needs a Cookie rule.
Above the form are the built-in groups:
| Group | What it does |
|---|---|
| WordPress | Skips logged-in visitors, wp-admin, logins, comments, files and the panel’s own checks. It cannot be switched off. |
| Phones | Keeps a separate copy for phones, by the same test as wp_is_mobile(). |
| PHP sessions | Skips a visitor with a PHP session. |
| WooCommerce | Skips once a visitor has a cart. Keeps a separate copy per ?v=. |
| Easy Digital Downloads | Skips once a visitor has a cart. |
| Currency switchers | Keeps a separate copy per currency, country or tax choice. |
| Languages | Keeps a separate copy per language cookie. |
| Tracking parameters | Ignores utm_*, gclid and the like. |
They stay on for every site, because a cookie rule matches nothing without its plugin. In use marks a group whose plugin the site has. Each group and each of its rules has a switch. To change a built-in value, switch that rule off and add your own. The rules you leave alone improve with each release.
Keep separate copies
Section titled “Keep separate copies”This kind of rule stores one copy of a page per value of a cookie or a parameter, such as a currency or a language.
- Set What to Separate copy per value.
- Set Matches to Cookie or Query parameter, and Value to its exact name.
- Choose Add rule.
A value has to be 1 to 32 letters, digits, _ or -. Any other value skips the cache, so a visitor
cannot fill it with made-up copies. Each rule multiplies the copies of every page, so a site has at
most five of its own.
Ignore a query parameter
Section titled “Ignore a query parameter”Ignore this parameter leaves a parameter out of the cache and takes it off the request PHP sees.
A link with it gets the stored page, and its value cannot end up inside one. A redirect keeps the
parameter. A * at the end of Value takes any name that starts so.
The Tracking parameters group does this for utm_*, gclid, gbraid, wbraid, fbclid,
msclkid, mc_cid, mc_eid, _ga and _gl. If the site reads one of them in PHP, switch that rule
off.
Keep a page for less time
Section titled “Keep a page for less time”Keep pages for sets the time for the whole site, from 5 minutes to 10 hours. Default takes it from Settings → Sites. Ten hours is the most, because WordPress’s form tokens expire after 12 to 24 hours.
For some pages only:
- Set What to Keep for a set time.
- Set Matches to URL path or Post type, fill in Value, and set Seconds.
- Choose Add rule.
A path rule comes first, then a post type rule, then the site’s time.
Check a URL
Section titled “Check a URL”In Check a URL, enter an address on the site, such as /shop/, and choose Check. The panel
asks the site’s nginx from inside its container, as a visitor with no cookies. One line says what
happened, with the status when it is not 200.
| The line | What it means |
|---|---|
Cached. |
The page is stored. |
Skipped: and a reason |
The request went past the cache. |
Not stored: and a reason |
The page came from PHP, and was not stored. |
An old copy: PHP or the database isn't answering. |
The last stored copy, because of a failure. |
The page cache is off. |
nginx runs without it, because it refused the panel’s file at a start. The site’s log says why. |
Every page also answers with X-Cache-Status, and with the reason in X-Cache-Skipped. HIT is a
stored page. MISS and EXPIRED came from PHP. STALE is the last stored copy, after a failure.
BYPASS went past the cache. A browser signed in to WordPress always gets BYPASS, because of its
cookie, and curl sends none:
curl -sI https://northwindbakery.example/ | grep -i x-cacheMost reasons say what they mean, such as logged in, page sets a cookie or
cookie wordpress_logged_in_. These need a word more:
| Reason | What it means |
|---|---|
query string |
A query parameter the cache does not know. |
file |
The address looks like a file. |
body |
A GET with a body. |
ajax |
A request from a script: X-Requested-With, X-PJAX, HX-Request, Turbo-Frame or Origin. |
range |
A request for part of an answer. |
device |
The browser’s phone hint and its user agent disagree. |
separate copy value |
A separate-copy cookie or parameter has a value that is not a short token. |
page says no-cache |
Cache-Control says no-cache, no-store, private or a zero age, or Expires has passed. |
page varies by … |
Vary names something the cache keeps no copies by. |
page flushed early |
The page was sent in parts, so it streams instead. |
missing file |
A 404 at an address with a file extension. |
another spelling of this address |
Such as /Shop/ for /shop/. |
request header … |
The request carries a header the cache does not know, usually from a proxy. |
.htaccess limits who can see this |
A .htaccess above WordPress says who may see the page. |
.htaccess or Apache says no-cache |
Something after PHP said so, such as a Header line. |
cleared while rendering |
A clear came while the page was being drawn. |
not a WordPress page |
Check a URL only: nothing asked to store this answer. |
Limits
Section titled “Limits”| Limit | Value |
|---|---|
| A site’s page cache | About 256 MB. The least used pages go first. |
| The longest a page stays | 10 hours |
| A 404 | 5 minutes |
| A site’s own rules | 50, of which 5 separate-copy rules |
| Redis | An eighth of the site’s memory limit, at least 16 MB, out of that limit |
| A key in Redis | 4 hours |
- Logged-in visitors never get a stored page. The object cache is what speeds them up.
- A stored page never reaches
.htaccessor PHP. A change to the root.htaccessempties the cache within 10 seconds. A block in PHP, such as a security plugin’s firewall, or in another folder’s.htaccess, does not apply to a stored page until it expires. - Site protection and blocked addresses always apply, because they come before the cache.
User-Agent,Accept,Accept-LanguageandX-Forwarded-*reach PHP but do not keep copies apart. A page that changes by them without saying so inVaryfreezes the first visitor’s version.- A page that changes by what the request does not show, such as random content or location by address, needs a skip rule or Off.
- A request with a cookie the cache does not know cannot store a page, though it can get one. Consent cookies count as such. On a site where every visitor has one, nothing is stored.
- Behind a CDN other than Cloudflare, a header the cache does not know can stop pages being stored.
X-Cache-Skippedon a page that came through the CDN names it. Check a URL does not go through the CDN, so it cannot. - A store’s country header works only from Cloudflare.
- A clear empties the whole site’s cache. There is no clear for one page.
- A visitor’s approved comment clears the cache at most once a minute.
- After a clear or an expiry, every request for a page goes to PHP until one has been stored. PHP’s workers take them in turn.
- A 404 at an address with no file extension stays for its 5 minutes, even if you upload a folder there over FTP. Clear cache ends it.
- Tracking parameters are not put back on a redirect to another domain, or when the address already has a query and the request has a separate-copy parameter.
- A site hosted elsewhere has no caches. Neither has a site that still runs the old PHP setup: see Updating WPL7.
- A PHP site has no page cache. Its Redis is for its own code. The panel never clears it, but every restart empties it.