How long can your site be offline before Google removes it from its index?

What does Google mean by "extended period"?

Google intentionally leaves the exact duration vague. Several days can mean anywhere from 48 hours to a week, depending on the context. The engine does not apply a fixed threshold but uses heuristics to assess each situation.

These heuristics take into account the site's reliability history, its usual crawl frequency, and likely the perceived importance of its content. A major news site will benefit from greater tolerance than a little-read blog.

How does Google decide to remove a site from the index?

The process relies on the accumulation of HTTP errors during successive crawl attempts. If Googlebot consistently encounters 500, 503 errors or timeouts, it interprets these signals as a lasting unavailability.

Unlike a 404 error that indicates a specific page no longer exists, a global server error suggests a significant technical problem. Google cannot know if the site will be back tomorrow or never. As a precaution, it eventually cleans its index of inaccessible URLs.

Why does this policy exist?

Google seeks to maintain the quality of its results. Providing links to down sites degrades user experience. The engine must balance between being patient with temporarily unavailable sites and quickly dismissing those that may never return.

This approach also protects crawl resources. Why continue to request a server that never responds? It is better to reallocate this budget to active sites.

• No fixed deadline: Google adjusts its decision based on the site's history and authority.
• Critical server errors: 500/503 codes trigger alerts faster than 404s.
• Quality objective: Google prioritizes user experience over infinite patience.
• Crawl optimization: avoid wasting resources on dead sites.
• Possible reversibility: a site that returns can be reindexed, but not instantly.

Do these statements align with real-world observations?

Yes, generally. We regularly observe mass deindexing following failed migrations or extended server outages. A site that remains in maintenance for 72 hours without a backup page often sees its rankings plummet.

But the devil is in the details. [To be verified] Google never specifies the exact threshold in hours, nor how domain authority affects this timeframe. A site with a high historical PageRank and strong backlinks will likely hold out longer than a new domain without notoriety.

What nuances should be considered?

Not all outages are equal. A 503 error with a correctly set Retry-After header signals a planned temporary maintenance. Google may interpret this signal differently than a sudden timeout or a generic 500 error.

Similarly, an outage affecting 100% of the site will have a more significant impact than a partial problem affecting certain sections. If Googlebot can successfully crawl 80% of your URLs but encounters sporadic errors on the remaining 20%, it will likely not deindex the entire domain. Context is crucial.

In what situations does this rule apply differently?

High-frequency news sites seem to have a paradoxical reduced tolerance. Google crawls these domains intensively and thus detects anomalies faster, but it also gives more credit to their quick return. A media site that is down for 6 hours may lose fresh rankings without being completely deindexed.

Conversely, a corporate site with a low update frequency may remain in maintenance for 48 hours without immediate consequences, since Google only visits it every three days. The usual crawl rate partly determines the engine's reaction speed.

What should be implemented before a planned maintenance?

Set up a clean maintenance page with a 503 HTTP code and a Retry-After header indicating the estimated duration. This informs Googlebot that the unavailability is temporary and planned, which may influence its response.

Keep at least a few critical URLs accessible: homepage, main category pages, high SEO value content. If the entire site cannot stay online, isolate strategic sections on backup infrastructure.

How to react if an unexpected outage occurs?

Restore service as quickly as possible, of course. But do not count on instant reindexing. Manually submit your priority URLs via the Search Console as soon as you are back online to accelerate the process.

Monitor your server logs and the Search Console to identify which sections Google attempted to crawl during the outage. This data indicates which URLs are most at risk of losing their indexed status. Focus your recovery efforts on these priorities.

What mistakes should be absolutely avoided?

Never let a site return silent 500 errors for days, hoping no one will notice. Google always notices. And your competitors too, who will take advantage to eat into your positions.

Also avoid switching to a generic 200 OK homepage displaying "Under Maintenance" on all URLs. Google will interpret this as massive duplicate content or thin content, which can trigger other algorithmic issues.

• Implement a 503 code with Retry-After during planned maintenance.
• Test your maintenance configuration on a subdomain before deployment.
• Keep at least the homepage and critical conversion pages accessible.
• Monitor Search Console daily during and after the incident.
• Manually submit priority URLs via XML sitemap as soon as back online.
• Document each outage with its duration and observed SEO impact to refine your protocols.