What «too many redirects» actually means
You open your site and the browser shows a strange message. Chrome displays ERR_TOO_MANY_REDIRECTS with «this page isn't working, it redirected too many times». Firefox says «the page isn't redirecting properly». Safari shows an equivalent warning. All three tell you the same thing.
This is not a browser bug: it is a safety stop. A redirect happens when a URL tells the browser «don't load this, go to that other address». The problem appears when that other URL points back to the first, or when a third one closes the circle. The browser enters a chain that never ends and, to avoid spinning forever, stops.
It is not a 500 error, a blank page, or an invalid HTTPS warning. In a loop, each hop usually works: the problem is the sum.
The warning alone does not mean you have lost content. It means the browser cannot reach an address that finishes loading. Before changing WordPress, identify which addresses keep sending visitors back and forth.
How the loop builds up: a chain that never ends
To understand the loop it helps to know how a redirect works. When you request a URL, the server can answer with a 3xx code (301, 302, 307 or 308) and a Location header saying where to go next. The browser follows; if it receives a 200, it renders. Another 3xx and it jumps again.
The full mechanics are in the MDN reference on HTTP redirects. The formal definition is in RFC 9110, 3xx section, and the header that governs each hop is explained in the MDN entry for Location. The Google Search Central guide on redirects covers how the search engine interprets a chain.
The classic loop appears when two URLs send each other back: /page redirects to /page/ and /page/ redirects to /page. Three other shapes are common:
- HTTP ↔ HTTPS loop. One rule forces HTTPS, another forces HTTP.
- With www ↔ without www loop. One place says
www.yourdomain.com; another saysyourdomain.com. - Main domain ↔ alias loop. An alias redirects to the main domain and an old rule sends it back.
Browsers limit the number of redirects before showing the warning; the limit varies. After correcting the cause, test in a private session and clear stored redirects from caches you manage. An old response can make the error appear to persist.
Common causes on a WordPress site
On WordPress, these are some configurations to compare with the redirect chain you actually observe:
- WordPress addresses that do not match the installation.
siteurlidentifies the WordPress files andhomethe public address. They can differ when WordPress is in a subdirectory. Conflicting redirects are the problem, not different values on their own. - HTTPS or SEO plugins with conflicting rules. One forces HTTPS, another sets a different canonical, another adds a hop for the trailing slash. The set bites its own tail.
- Leftover rules in
.htaccess. On Apache (the usual server on shared hosting) rewrite rules live there. After migrations or old plugins, fragments remain that should not. - Cloudflare’s Flexible mode. Cloudflare can distribute content before it reaches the visitor. In Flexible mode, it connects to your server over HTTP. If that server redirects HTTP requests to HTTPS, a loop may result.
- Reverse proxy with its own redirect. A reverse proxy (a server in front of the real one that forwards requests) adds its own hop towards the canonical.
- Page cache serving a stale version. A plugin saved the response when the origin was broken and keeps returning it.
WordPress is rarely the only culprit. The cause usually sits between layers (hosting, DNS, CDN, WordPress), when two of them decide different things about which URL is right.
Before touching anything: check and write it down
Start with checks that do not change the site. If redirect rules need editing, prepare a recoverable backup and coordinate with whoever maintains the website.
- Test in a private window, no extensions, no session signed in. Rules out stale browser data.
- Test the four combinations:
http://yourdomain.com,https://yourdomain.com,http://www.yourdomain.com,https://www.yourdomain.com. Note what each returns and, if it redirects, where. - If you use browser developer tools, open Network and preserve the log while reloading. The chain shows addresses, response codes and the
Locationheader pointing to the next address. Your team can also inspect headers withcurl -I https://yourdomain.com. Do not disable certificate checks to hide a separate problem. - Check the scope. Only the home or also
/wp-admin/? Whole site or a section? - Test from another network (mobile data). If it works there, the loop may be in a local cache.
- Save screenshots and the list of URLs. You will need them to decide and, if you ask for help, to explain what you saw.
These checks let you explain the problem without sharing passwords: affected address, time, browser and redirect chain. You do not need to disable plugins to collect that information.
Hypothetical example: the HTTPS address without www redirects to HTTP with www, which redirects back to the original address. That repeating pair locates the loop. The next step is to identify which setting creates each redirect.
Domain with and without www, HTTP and HTTPS: the most common loop
Check HTTP/HTTPS and www/non-www versions of the same address. The site should have a preferred public version, and the others should reach it without returning to the start. Avoid unnecessary intermediate redirects where possible.
The classic trap: a WordPress plugin forces HTTPS, the server forces www because of an old rule. If you enter via https://yourdomain.com, the server sends you to http://www.yourdomain.com to add the www, and the plugin sends you back to https://yourdomain.com to force HTTPS. Each rule does what you asked; together, a loop.
| Combination | What it should return | How to check |
|---|---|---|
| Canonical URL | A direct 200 response with your content. | curl -I returns a 200 status, regardless of the HTTP version, with no Location. |
| Variant with the other protocol | A permanent redirect, usually 301 or 308, to the preferred address. | curl -I returns 301 or 308 and a Location pointing to the canonical. |
| Variant with the other «www» | A permanent redirect, usually 301 or 308, to the preferred address. | One hop, one Location pointing to the canonical. |
| Variant that mixes both | A permanent redirect, usually 301 or 308, to the preferred address. | One hop, not two. A 301 pointing back to the starting URL is your loop. |
Choose a preferred public version and avoid unnecessary intermediate redirects. A direct redirect is a useful target, but two redirects do not themselves make a loop. Avoid a rule returning to an address already visited.
If the loop also hits the admin, the official WordPress documentation on changing the Site URL describes how to do it from the database or from wp-config.php. It is a delicate change; without backing up the file or the wp_options table first, you may complicate the diagnosis instead of fixing it.
Site URL and Home URL: where they live and how to review them safely
WordPress stores two key URLs in wp_options: siteurl (where the core lives) and home (the address visitors see). They can also be forced from wp-config.php with WP_SITEURL and WP_HOME, which take precedence over the database.
If the admin opens, go to Settings → General. If it does not, three alternatives:
- Read
wp-config.phpover FTP or from the hosting panel. Look fordefine('WP_HOME', ...)anddefine('WP_SITEURL', ...). If they exist, their value is the effective one. - phpMyAdmin. In
wp_options, checksiteurlandhome. Before changing, export the table or write the current values down. - WP-CLI.
wp option get siteurlreads without touching; for changes,wp option update, same precaution.
Each value must match the location it describes. If the WordPress files are in a subdirectory, siteurl and home can legitimately differ. Do not make them identical blindly: check the domain, protocol, directories and how the server recognises HTTPS connections.
One detail that saves trouble: if you change these values, purge the cache. Many cache plugins store absolute URLs in the cached HTML. Purge the plugin caches and the CDN, in that order. Otherwise the old response keeps being served and the loop appears to be still there.
.htaccess, CDN and proxy: the layers that also redirect
Above WordPress, several layers can add their own redirects. When the loop survives after reviewing Site URL and Home URL, it is often in one of them.
- .htaccess. On Apache it lives at the web root. After migrations or years of plugins, it is common to find duplicated blocks, rules that force HTTPS alongside others that force HTTP, or redirects inherited from a previous domain. Before deleting, copy the file; sometimes the issue is the order between rules, not a single rule.
- Nginx. If your host uses Nginx, the rules live in server configuration (
serverandlocationblocks). Typically only the host team has access; ask them for the review. - A CDN such as Cloudflare. In Flexible mode, Cloudflare connects to the origin over HTTP. If the origin forces HTTPS, a loop can result. With a valid, supported origin certificate, coordinate a change to Full (strict), which verifies that certificate too. Full encrypts the connection without the same verification. Review redirect rules as well; do not change the mode without checking the origin.
- Hosting-level rules. Some panels include a screen to force the canonical version. If it is active and WordPress forces the opposite, loop guaranteed.
Your team can compare the CDN response with the origin response while preserving the hostname and HTTPS verification and accounting for origin access restrictions. A difference helps locate the layer involved, but does not by itself prove which rule is wrong.
We describe this layered architecture in the piece on HTTPS, WAF and CDN; and if the loop appeared after a move, the website migration and SEO checks guide reviews the residual rules a poorly closed migration drags along.
Plugins, themes and cache: check them cold
If server and CDN rules do not explain the loop, your team can investigate plugin redirects, access restrictions and theme code. A page might send logged-out visitors to the login screen, which then sends them back to that page.
Test the hypothesis on an isolated staging copy with protected data and live payments, emails and integrations disabled. Disable only the suspected component, test the same addresses and restore its state before moving on. Renaming a folder on the live website can also remove important functions: it is not a harmless check.
A staging result guides the fix, but behaviour can differ if hosting or CDN settings differ. Coordinate the final change, keep a rollback route and check public pages and login. Purge only the affected caches.
If you want us to walk you through it
A redirect loop is easier to diagnose with someone looking at the different layers at the same time: hosting, DNS, CDN, WordPress. If you have worked through the steps and the error persists, tell us what you have seen: the browser, the URLs you tried, the hops from the Network tab. We will review it and explain the proposed work before starting.
If your goal is not a specific loop but preventing it from showing up again, reviewing the site calmly makes sense. You can start with the piece on HTTPS, WAF and CDN to see how those layers interact in this kind of error, and in our performance area you will find the scope where we place these reviews.