Does Downtime Hurt SEO? What Really Happens to Rankings

Does downtime hurt SEO? A short outage almost never does, but downtime that lasts for days can reduce how often Google crawls a site and eventually push pages out of the index. What decides the outcome is how long the site is unavailable and which HTTP status code it returns while it's down.

For an agency, fixing the outage is usually someone else's job: the host, the client's developer, their IT team. Noticing it is yours. A site Googlebot can't reach can't hold its rankings, and the client will ask why nobody caught it.

What Google's documentation says

Google's page on how HTTP status codes affect crawling is fairly direct. 5xx server errors and 429 Too Many Requests make Googlebot slow down its crawling of the site. URLs that are already indexed are kept for a while, but if the errors continue, they are eventually dropped. Timeouts and refused connections are handled much like server errors.

For planned or temporary unavailability, 503 Service Unavailable is the recommended response. It tells crawlers the problem is expected to pass. A Retry-After header adds when to come back, for example Retry-After: 3600 for one hour.

One detail is easy to miss. If robots.txt itself returns a server error, Google's robots.txt documentation says it stops crawling the site, and if the error lasts a long time it falls back to the last cached copy of the file. An outage that takes down the whole host takes robots.txt with it.

How long is too long

Google publishes no cutoff, so read these as rough bands rather than rules.

Outage length	What usually happens
A few minutes	Googlebot may not request anything. If it does, it retries later. No measurable ranking effect.
A few hours	Crawling slows. Pages stay indexed, and with a clean 503 the site recovers once it's stable.
One or more days of 5xx	Crawl rate drops and some URLs can fall out of the index. Most return after recovery, not always in the same position.
Frequent short outages	Each is minor, but a host that fails repeatedly can keep Googlebot crawling cautiously.

Why the status code matters as much as the duration

Two sites down for the same four hours can come out of it differently.

The first returns 503 with Retry-After from its maintenance mode. Google reads that as temporary and waits. The second shows a friendly "We'll be back soon" page with 200 OK. To Googlebot that is a real page whose content has been replaced by one sentence, and if it's crawled in that state the maintenance text can end up in the search snippet until the real content is back.

A 500 Internal Server Error or a timeout sits between the two. It won't be indexed as content, but it doesn't carry the "this is temporary" signal either.

WordPress gets this right by default. During core and plugin updates it creates a .maintenance file and serves "Briefly unavailable for scheduled maintenance" with a 503 status. Third-party "coming soon" and maintenance plugins vary, so check which one a client's site uses before a planned outage.

Where the damage shows up in Search Console

If you suspect an outage has already affected a client, two reports help. The Crawl stats report, under Settings, has a Host status section covering robots.txt fetching, DNS resolution and server connectivity over the last 90 days, plus a breakdown of crawl requests by response code. The Page indexing report lists URLs under "Server error (5xx)" when Google failed to fetch them.

Both look backwards. They explain what happened last week, which is useful for a client report and no help in stopping it. For another infrastructure failure that looks like downtime to visitors, see SSL certificate expiry monitoring.

Finding out while it is still happening

Picture the expensive version. A client site goes down late on a Friday, nobody is looking, Googlebot and customers hit errors all weekend, and you hear about it on Monday. For an ecommerce client the lost weekend sales may matter more than any ranking dip, and you're the one explaining both. Checking 5 or 50 client sites by hand isn't realistic, so the workable answer is a monitor that alerts on failure. Our UptimeRobot alternative comparison walks through the options.

Deltio's uptime monitoring alerts you on Slack and email, per site, when a client site goes down. It also detects when a page has been swapped for a maintenance page, which catches the 200 OK "back soon" case that a plain up/down check reports as healthy. Because the same account compares noindex, robots.txt and sitemap URLs on a daily cycle, you also see if a site comes back from an incident with its SEO settings changed.

It doesn't analyse crawl budget or read server logs. For those you still need Search Console and the host's access logs. How uptime and indexability fit together is covered in uptime and SEO monitoring.