How long to keep 301 redirects after a site move
How long to keep 301 redirects: Google says as long as possible, generally at least a year. What to check in your logs and links before removing any.
By Crawlito · 9 min read
In short
- Google's site move guide says to keep redirects for as long as possible, generally at least 1 year, and to consider keeping them indefinitely for users.
- Google's Change of Address help says to maintain redirects for at least 180 days, longer if they still get Google Search traffic, and to keep paying for the old domain for at least a year.
- Google says the one-year timeframe lets it transfer all signals to the new URLs, including recrawling and reassigning links on other sites that point to the old ones.
- Before removing a redirect, update internal links, sitemaps and canonical tags, then check your logs for crawler and referral hits on the old URLs.
- When a second migration stacks on the first, point every old URL straight at the newest address instead of adding another hop.
On this page
- What Google says about how long to keep redirects
- Why the figure is a year, not a few weeks
- What keeping redirects actually costs
- What to check before removing any redirect
- How to check your logs before removing redirects
- What happens if you remove redirects too early
- Frequently asked questions
- Check your redirects end to end
Keep 301 redirects for as long as possible, and for at least one year after a site move. That is the figure in Google's guide to moving a site with URL changes, which also says that from users' perspective you should consider keeping redirects indefinitely. In practice, a redirect is cheap to keep and expensive to lose, so the real question is not "when can I delete these?" but "is anything still arriving at the old URLs?". Your server logs answer that.
This post covers what Google's documents say about duration, the reasoning behind the numbers, what to clean up before removing anything, and a log-based check you can run. For planning the migration itself, see preserving discovery during a website migration. For how crawlers read each status code, see status codes and redirects for AI crawlers.
What Google says about how long to keep redirects
Three Google documents touch on the question, and they give different numbers for different things. Read together, they are consistent.
| Source | What it says | Applies to |
|---|---|---|
| Site moves with URL changes | Keep redirects as long as possible, generally at least 1 year; for users, consider keeping them indefinitely | Any move where URLs change |
| Change of Address tool help | Maintain redirects for at least 180 days, and longer if they still get Google Search traffic | Domain or subdomain moves submitted in Search Console |
| Change of Address tool help | Keep paying for the old domain for at least a year | Domain moves |
| Redirects and Google Search | Use a permanent redirect when you are sure it won't be reverted; no duration given | All redirects |
The 180 days in the Change of Address help is a floor tied to that tool: Google says its migration signals continue for 180 days after you start the move in Search Console. The same page says that after the 180-day period Google no longer recognizes any relationship between the old and new sites, which makes the redirects the only remaining link between them. It is not a recommendation to delete the redirects on day 181. The site move guide's "at least 1 year" is the general figure, and "as long as possible" is the instruction.
Other operators say less. At the time of writing, OpenAI, Anthropic and Perplexity do not document how their crawlers handle redirects or how long they keep old URLs, so there is no AI-specific timeline to follow. Google's figure is the best documented one available.
Why the figure is a year, not a few weeks
A site move looks finished long before it is. Google's site move guide says that to consider a move complete, Googlebot has to visit every URL on both the old and new site at least once, and that a small to medium-sized site can take a few weeks for most pages to move, with larger sites taking longer. Because every URL needs at least one visit, pages that Googlebot seldom crawls keep the move open.
The year covers more than Google's own crawl. The guide gives the reason directly: the timeframe allows Google to transfer all signals to the new URLs, including recrawling and reassigning links on other sites that point to your old URLs. You do not control when Google revisits a page on someone else's site that links to you, and every one of those links still names the old address.
Several things outlast any crawl cycle:
- Links on other sites. Articles, forum posts and directories written years ago keep pointing at old URLs unless their owners edit them.
- Bookmarks, emails and printed material. None of them update themselves.
- Copies of your old URLs held by AI tools and search indexes. An assistant may still have an old address from a past crawl and fetch it when a user asks a question. No operator documents when those references expire.
- Browser caches. RFC 9110 says 301 and 308 responses are heuristically cacheable, so a returning visitor's browser may skip your server and follow the cached redirect. That also means a wrong permanent redirect can stick for some users after you fix it.
What keeping redirects actually costs
For most sites, very little in server work. The nginx map documentation, for example, says exact strings are looked up in hash tables and that map variables are evaluated only when used, so declaring a large number of them adds no extra cost to request processing. The real costs are elsewhere:
- Chains. Each migration adds a layer. If the 2022 move sent
/ato/band the 2025 move sent/bto/c, every visit to/anow takes two hops. - Rules nobody understands. Regex redirects written for an old platform can catch new URLs by accident.
- Speed for real people. Google's site move guide notes that redirects are slow for users, which is why it asks you to update your own links rather than lean on redirects forever.
The fix for all three is to maintain the redirects, not delete them. Google's guide says to avoid chaining redirects and redirect to the final destination directly; where that is not possible, keep chains short, ideally no more than 3 and fewer than 5 hops. When a second migration stacks on the first, rewrite the old rules so every historical URL points straight at its current address:
# One lookup table for every migration, old path -> current path.
# map belongs in the http context. $uri is the path without the query
# string; plain string keys match exactly, ignoring case.
map $uri $redirect_target {
default "";
/a /c; # 2022 move, updated in 2025
/b /c; # 2025 move
/old-blog/pricing /pricing;
}
server {
listen 443 ssl;
server_name www.example.com;
if ($redirect_target) {
return 301 https://www.example.com$redirect_target;
}
# ... rest of the site
}Keep the mapping file in version control with a comment for each batch, so the next team knows why a rule exists. For choosing between 301 and 308, see 301 vs 302 vs 307 vs 308. If Search Console already reports a problem, Redirect error in Search Console lists the four causes Google names.
What to check before removing any redirect
If you do want to retire some rules, clear these first. Each one removes a reason for traffic to keep arriving at the old URLs.
- Internal links. Google's guide says to replace all internal links that point to old URLs, using your URL mapping. Search templates, navigation, body content, structured data and downloadable files for the old host and path patterns.
- Sitemaps. Your sitemap should list only new URLs. Google says you can remove the old sitemap once it is using the new one going forward.
- Canonical and hreflang tags. Google's guide says each new URL should have a self-referencing
rel="canonical"tag, and that hreflang annotations should be updated to use the new URLs. A canonical that still names an old URL sends crawlers back through the redirect. - High-volume external links. The guide suggests contacting sites that link to your old content, prioritizing by the number of visits each link sends. One updated link from a busy page can retire a lot of redirect traffic.
- The old domain. For a domain move, Google's Change of Address help recommends paying for the old domain for at least a year so that nobody else can register it and use it maliciously. If the domain lapses, every redirect on it goes with it, and someone else can serve whatever they like at your old addresses.
- A Change of Address request still in its window. Google says you can cancel a request for 180 days, and that if you do, you should remove the 301s first because Google may keep following them. Do not remove redirects as a side effect of other work during that period without knowing which outcome you want.
How to check your logs before removing redirects
Logs show who still uses each old URL. Run the check over as long a window as your log retention allows, ideally several months, because some crawlers and referrers visit rarely. The commands below assume the standard nginx or Apache combined log format, where the request, referrer and user agent sit in double quotes.
Count which user agents still hit redirecting URLs:
zcat -f /var/log/nginx/access.log* \
| awk -F'"' '$3 ~ /^ 30[18] / {print $6}' \
| sort | uniq -c | sort -rn | head -20List the old paths that still get requests, most used first:
zcat -f /var/log/nginx/access.log* \
| awk -F'"' '$3 ~ /^ 30[18] / {split($2, r, " "); print r[2]}' \
| sort | uniq -c | sort -rn | head -50Find the external pages that still send people to old URLs:
zcat -f /var/log/nginx/access.log* \
| awk -F'"' '$3 ~ /^ 30[18] / && $4 != "-" && $4 !~ /example\.com/ {print $4}' \
| sort | uniq -c | sort -rn | head -30Read the results like this:
| What you see | What it means | What to do |
|---|---|---|
| Googlebot, Bingbot or AI crawlers on old URLs | Those systems still hold the old addresses | Keep the redirects; confirm the crawler is real before trusting the count |
| External referrers on old URLs | A live page still links to the old address | Ask the site owner to update the link, and keep the redirect meanwhile |
| Browser user agents with no referrer | Bookmarks, typed URLs, email or cached links | Keep the redirect; you cannot update these |
| No hits at all over several months | Nothing currently depends on the rule | A candidate for removal, if you still want to remove it |
User agents can be faked, so check the IPs of crawler hits against each operator's published ranges before acting on them. AI crawlers in server logs shows how to count and verify those visits. One caveat: if your CDN answers redirects at the edge, the request may never reach your origin, so run the same check on the CDN's logs instead.
There is no documented threshold for "few enough hits". That decision is yours, and given how cheap a redirect rule is, a handful of real visits a month is usually reason enough to keep it.
What happens if you remove redirects too early
When a redirect disappears, the old URL starts returning whatever the server now has at that path, usually a 404. A crawler that still had the old address gets an error instead of your page, and a visitor following an old link lands on a dead end. Google's redirects documentation notes that after a domain move, Google is very likely to keep occasionally showing old URLs in results even after the new ones are indexed, and that this is normal. Removing the redirect during that period turns those results into errors.
If you are removing redirects because content is truly gone, that is a different case: Google's site move guide says URLs for content that is not moving should return 404 or 410. Retire the content deliberately rather than as a side effect of a redirect cleanup.
Frequently asked questions
Is 180 days enough to keep redirects after a domain change?
Only as a minimum. Google's Change of Address help says at least 180 days and longer if the redirects still get Google Search traffic, while its site move guide says as long as possible, generally at least 1 year.
Do I need to keep the old domain after redirecting it?
Yes, for at least a year according to Google's Change of Address help, and the redirects only work while you own it. If the domain lapses, someone else can register it and serve their own content at your old URLs.
Can I delete the redirects once Google shows the new URLs?
Not safely. Google showing the new URLs does not mean every link on other sites, every bookmark and every AI tool's stored address has moved. Check your logs for remaining hits first.
Should old redirects point to the newest URL after a second migration?
Yes. Google says to avoid chains and redirect to the final destination directly, so update the first migration's rules to point at the current address instead of adding a second hop.
Do AI crawlers need redirects kept longer than Google does?
No operator documents this at the time of writing. AI tools can hold old URLs from past crawls, so the same rule applies: keep redirects while your logs show anything still using them.
Check your redirects end to end
Before and after any cleanup, test a sample of old URLs and confirm each one reaches its final page in one hop with a 200. The free check shows the status code and redirect path for a single page as AI crawlers would fetch it, and an audit covers redirects, sitemaps and canonical tags across up to 500 pages with the evidence for each fix.
Filed under Access and robots.txt. Topics: redirects, migration, google, server-logs