Managing 10,000+ 301 Redirects: What Breaks and What Scales

Ten thousand redirects arrive faster than most teams expect: one replatforming of an online store, two rebrands, a few years of campaign URLs. The redirects themselves are simple. What changes at that size is speed, who can change them, and whether anyone can still tell which ones matter.

What breaks after 10,000 redirects#

• Speed. Rule lists are checked top to bottom. Ten thousand RewriteRule lines in .htaccess are read and tested on every request, including requests for pages that don't redirect at all.
• Chains. Each migration adds redirects on top of the last one, so old URLs pass through two or three hops. Google's crawlers follow up to 10 redirect hops by default^[1], but every hop is a full round trip for visitors.
• Ownership. When redirects live in a server file, only engineers can change them, and marketing or SEO waits for a deploy.
• Auditing. Nobody remembers why a redirect exists, whether it still gets traffic, or who changed it last.

The options compared#

Cloudflare's Bulk Redirects allow 10,000 URL redirects across lists on Free, 25,000 on Pro, 50,000 on Business and 1,000,000 by default on Enterprise, per account^[2]. Single Redirect rules are a separate, much smaller allowance per zone.

Nginx map example#

nginx# /etc/nginx/redirects.map  (one line per redirect)
# /old-path  /new-path;

map $uri $redirect_to {
    default "";
    include /etc/nginx/redirects.map;
}

server {
    server_name www.example.com;
    if ($redirect_to) {
        return 301 $redirect_to;
    }
}

Apache RewriteMap example#

apache# In the virtual host config, not .htaccess
RewriteEngine On
RewriteMap redirects "dbm:/etc/apache2/redirects.map"
RewriteCond ${redirects:%{REQUEST_URI}} !=""
RewriteRule ^ ${redirects:%{REQUEST_URI}} [R=301,L]

# Build the DBM file from a text file of "old new" pairs:
# httxt2dbm -i redirects.txt -o redirects.map

Which one to choose#

Keeping a large redirect list healthy#

1. Flatten chains after every migration. Point every old URL straight at the final URL, not at the previous redirect. See Fix redirect chains and loops.
2. Prefer exact matches. Use pattern rules for whole folders that moved the same way, and exact entries for everything else. Exact lookups stay fast and are easy to read.
3. Tag every redirect with its project. "2024 replatform" or "Spring campaign" tells the next person why it exists.
4. Review the ones nobody visits. Keep redirects for moved pages at least a year. After that, look at redirects with no visits and no external links before removing them; removing a redirect that still has links turns it into a 404.
5. Record who changed what. Git history for map files, or change history in your redirect tool. It turns a broken redirect from a mystery into a two-minute fix.
6. Test a sample after each import. Check 50 old URLs from the list: each should return one 301 to a page that answers 200.

How RedirHub handles large redirect lists#

RedirHub answers redirects from 130 Global PoPs, so the lookup happens at the edge and your own server isn't involved. For large lists:

• Import from a CSV with a review of what will be added, updated or skipped before anything is written. Re-importing an edited export only counts new links.
• Capacity: Core includes 2,500 managed links, Pro 10,000, and Enterprise is set by contract for millions-scale operations. A whole-domain redirect uses no managed links.
• On Pro, every link shows who changed it, when, and the old and new values, including changes from CSV imports and the API.
• Change many destinations at once with a preview, from the dashboard or the API, or keep the list in Git with the Terraform provider.