How should temporary site maintenance be signaled to Google?

Definition: As of October 2026, Google recommends returning an informational error page with the HTTP status code 503 Service Unavailable when a website must close briefly. HTTP status codes describe the outcome of a server's response; 503 identifies temporary overload or scheduled maintenance.

Key facts

• Google's website pause guidance recommends a 503 error page for an urgent shutdown lasting one or two days.
• The HTTP semantics standard defines Retry-After as a response header indicating how long a client ought to wait before another request.
• Google's HTTP status documentation says 5xx server errors and 429 responses temporarily slow crawling.
• Google's network error documentation says already indexed URLs that remain unreachable are removed from its index within days.

What should a maintenance response contain?

Google's shutdown instructions pair the 503 response with a Retry-After header containing a best effort date or duration. The error page should use static HTML and minimize external resources. Its content can include an expected reopening date or customer service contact details.

The header's syntax comes from RFC 9110, section 10.2.3. Its value is either an HTTP date or a non-negative integer specifying seconds after receipt of the response. This illustrative response combines Google's maintenance status with the standard's sample delay:

HTTP/1.1 503 Service Unavailable
Retry-After: 120

Here, 120 means a delay of two minutes. The standard describes the header with 503 as the expected period of service unavailability.

What happens to crawling and indexing?

Google's HTTP status documentation says existing indexed URLs initially remain indexed during server errors, but persistent errors lead to removal. Google ignores content returned with 5xx responses. When the server resumes successful 2xx responses, crawling increases gradually.

Google's network troubleshooting documentation treats network timeouts, connection resets and domain name system (DNS) errors similarly to server errors: crawling starts slowing immediately. These failures can occur before the server sends a status code. Google recommends examining firewall rules and logs when diagnosing timeouts or connection resets.

How can the response be checked?

Google's maintenance checklist gives this command for confirming the response locally, using an example domain:

curl -I -X GET "https://www.example.com/"

The expected maintenance status is 503 Service Unavailable. The same checklist says to continue allowing crawling through robots.txt, the file containing crawler access rules. That file must remain accessible: returning 503 for robots.txt blocks all crawling. A rule disallowing all crawling can also cause content, and potentially URLs, to be removed from Google Search.

Related terms

• HTTP status codes
• Retry-After
• robots.txt
• Googlebot

FAQ

Should I use an access error or noindex during maintenance?

Google's website pause guidance warns against returning 403, 404 or 410, or applying a noindex instruction, to close a website temporarily. These choices remove the website's URLs from Google Search.

What if I need to close the site for longer?

The same guidance recommends keeping the site online with limited functionality where possible. If you need a longer shutdown, Google recommends an indexable placeholder home page returning 200.

Sources

1. Google, "How HTTP status codes affect Google's crawlers", updated 4 February 2026
2. Google Search Central, "Temporarily pause or disable a website", updated 10 December 2025
3. IETF, "RFC 9110: HTTP Semantics", June 2022
4. Google, "Debug network and DNS errors for Google's crawlers", updated 18 December 2025

---

Check how a page reads to answer engines. Open the GEO score analyzer