503 Service Unavailable

TL;DR — A 503 Service Unavailable503 Service Unavailable is the HTTP status code a server returns when it's temporarily unable to handle a request — usually because of overload or planned maintenance. It tells crawlers 'come back later' instead of 'this page is gone,' which is why Google recommends it (paired with a Retry-After header) for short, planned downtime. response tells search engines “the site is temporarily down — check back soon,” instead of “this page is gone.” When you need to take your site offline for planned maintenance, a 503 is the right code to return. Add a Retry-After header so GooglebotGooglebot is Google's web crawler — the software that fetches pages so Google can index and rank them. It comes in two variants, Googlebot Smartphone (primary, under mobile-first indexing) and Googlebot Desktop, and runs an evergreen Chromium renderer. knows when to come back, keep your robots.txt reachable, and don’t leave the 503 up for more than a day or two.

What a 503 actually means

HTTP 503 means the server is temporarily unable to handle the request, and the standard permits a Retry-After header. Evidence for this claim Primary standard or official documentation supporting the adjacent article claim. Scope: Protocol semantics and Search behavior are kept separate; no indexing, ranking, or migration-timing guarantee is inferred. Confidence: high · Verified: 2026-07-14 RFC 9110: 503 Google recommends 503 for brief downtime but warns that extended unavailability can affect indexingStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed.. Evidence for this claim Primary standard or official documentation supporting the adjacent article claim. Scope: Protocol semantics and Search behavior are kept separate; no indexing, ranking, or migration-timing guarantee is inferred. Confidence: high · Verified: 2026-07-14 Google: Pause an online business

A 503 is an HTTP status codeAn HTTP status code is the three-digit number a server returns with every response to tell a browser or crawler what happened to its request — success, redirect, client error, or server error. For SEO the code matters as much as the content: it tells Google and Bing whether to index a page, follow a redirect, retry later, or drop the URL from the index. in the 5xx family — the “something’s wrong on the server side” group. Specifically, it means the server is up enough to answer, but it can’t serve the request right now, usually because it’s overloaded or you’ve put it into maintenance mode.

The key word is temporary. That’s what separates a 503 from the other codes you might reach for when a page isn’t available:

• 200 OKHTTP 200 OK is the standard 2xx success status code, meaning the server received, understood, and fulfilled the request and is returning the resource. It's the code every page you want indexed should return — but a 200 alone doesn't guarantee Google will index the page. says “here’s the content” — even if that “content” is a “we’ll be right back” message. Search engines take that literally and can index the error page.
• 404 Not Found404 Not Found is the HTTP client-error status code a server returns when it can't find the requested URL — RFC 9110 defines it as no current representation, or unwillingness to disclose one. A \"hard 404\" actually returns the 404 status; a \"soft 404\" returns a success code (like 200) for a page that's really gone. 404s are normal and expected: the fact that some URLs 404 doesn't affect your site's other, successful pages, and Google de-indexes 404'd URLs over time (probably retrying for some period, less and less often). / 410 GoneA 410 Gone status code tells search engines a page was intentionally and permanently removed, prompting slightly faster de-indexing than a standard 404. say “this page doesn’t exist / is permanently gone.” Use these during maintenance and you’re telling Google to drop the page.
• 503 Service Unavailable says “I exist, I’m just busy — come back later.” This is the one you want for planned downtime.

Why 503 is the right call for maintenance

When you’re doing a deploy, a server move, or scheduled maintenance, you want search engines to pause, not to conclude your pages are dead. A 503 buys you that pause. Google says so directly: if you need to take a site down briefly, return a 503, not a 404 and not a 200 error page.

Think of it like a “Back in 10 minutes” sign on a shop door. A 404 is like bulldozing the shop; a 200 “coming soon” page is like replacing your whole inventory with a single sign and hoping customers still think you sell shoes. The 503 is the sign that keeps your spot.

The Retry-After header

Alongside the 503, you can send a Retry-After header that tells crawlersA crawler — also called a spider or bot — is an automated program that fetches web pages, extracts their links, and queues new URLs to visit. Search engines use crawlers to discover and download content for their index. roughly when to return. It can be a number of seconds or a specific date/time:

HTTP/1.1 503 Service Unavailable
Retry-After: 3600

That example asks bots to wait about an hour. Google may use it as a hint for when to recrawlCrawl frequency is how often a search engine comes back to re-fetch a page it already knows about. Popular pages that change often get refreshed many times a day; stable pages can go weeks or months between crawls — and you influence it indirectly, not by setting a dial.. It won’t come back to the exact second, but it won’t come back before then either.

The three things not to get wrong

1. Don’t 503 for long. A day or two is fine. Much longer and Google starts to think the outage is permanent and can drop your pages.
2. Don’t 503 your robots.txt file. If that file returns a 503, Google can’t crawl anything — including the check that would tell it you’re back online.
3. Don’t use a 404 or a plain 200 “coming soon” page instead. Those send the wrong signal and can cost you more than the 503 would.

Want the duration thresholds, the exact Google quotes, and the platform-by-platform implementation? Switch to the Advanced tab.

TL;DR — 503 is the correct code for temporary unavailability, and Google explicitly prefers it over 404/410 (permanent) or a 200 error page (indexable junk). Pair it with Retry-After. Google’s own ceiling for a full-site 503 is 1–2 days; past “a couple of days” it treats the 503 as permanent and drops URLs. A 503 also freezes your indexedStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed. metadata — titles, descriptions, and structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. won’t refresh while it’s served. Never 503 your robots.txt. Google treats 429429 Too Many Requests is an HTTP client-error status code that means a client sent too many requests in a given time window — a mechanism called rate limiting. Unlike other 4xx codes, Google treats 429 as a server-overload signal and slows crawling instead of removing content. and 503 the same for crawl-rate backoff. And recovery from a long outage is likely but not guaranteed to be 1:1. For anything longer than a day or two, serve an indexable 200 placeholder instead.

What 503 is — and isn’t

A 503 is a temporary server condition, not a permanent removal signal. Evidence for this claim Primary standard or official documentation supporting the adjacent article claim. Scope: Protocol semantics and Search behavior are kept separate; no indexing, ranking, or migration-timing guarantee is inferred. Confidence: high · Verified: 2026-07-14 RFC 9110: 503 Search behavior depends on duration and repeated responses; recovery timing is not guaranteed. Evidence for this claim Primary standard or official documentation supporting the adjacent article claim. Scope: Protocol semantics and Search behavior are kept separate; no indexing, ranking, or migration-timing guarantee is inferred. Confidence: high · Verified: 2026-07-14 Google: Pause an online business

RFC 9110 keeps the language soft on purpose: the server is currently unable to handle the request, relief is described as likely, not certain, and the Retry-After header is optional — a MAY, not a requirement. The standard doesn’t compel a crawlerA crawler — also called a spider or bot — is an automated program that fetches web pages, extracts their links, and queues new URLs to visit. Search engines use crawlers to discover and download content for their index. to return at any particular time, and a server can even skip the 503 entirely and just refuse the connection instead. That matters for this whole page: everything below about Google’s crawl behavior and recovery timing describes what tends to happen, not a contractual guarantee.

503 is for temporary server-side unavailability. It is not a general-purpose “hide this page” tool, and it is not interchangeable with its neighbors:

• 404 / 410 — permanent removal. During maintenance these tell Google your pages are gone, and it will start dropping them (410 a touch faster than 404).
• 200 with error content — a “we’ll be right back” page returning 200 OK looks like real content. Google indexes it, and if every URL returns the same page, it may treat them as duplicates.
• 503 — “I’m here, just temporarily unable.” The one code that says pause, not delete.

What Google actually recommends (in order)

Google’s current, actively-maintained guidance (Temporarily pause or disable a website) lays out a preference order that most competitor articles skip past. Lead with it:

1. Don’t go fully dark at all — limit functionality instead. Google’s top recommendation is to keep the site live and just disable the risky parts (turn off the cart, show a banner, update your structured data / Merchant CenterGoogle Merchant Center (GMC) is a free platform where retailers upload and manage product data so their products can appear across Google — Shopping, organic Search product grids, Images, Lens, and AI surfaces. Since 2020 it powers free (organic) product listings, not just paid Shopping ads. feed), because that “minimizes any negative effects on your site’s presence in Search.” Full-site 503 is the fallback, not the default.

2. If you must disable the whole site: the 1–2 day rule. Google calls a full-site takedown “an extreme measure that should only be taken for a very short period of time (a few days at most),” and is specific about the mechanism: “If you need to urgently disable the site for 1-2 days, then return an informational error page with a 503 HTTP response status code instead of all content.”

3. Longer than that: switch to an indexable 200 placeholder. Once you’re past a day or two, stop serving 503s. Google: “If you need to disable the site for a longer time, then provide an indexable home page as a placeholder for users to find in Search by using the 200 HTTP status codeAn HTTP status code is the three-digit number a server returns with every response to tell a browser or crawler what happened to its request — success, redirect, client error, or server error. For SEO the code matters as much as the content: it tells Google and Bing whether to index a page, follow a redirect, retry later, or drop the URL from the index..”

The duration spectrum (there’s no single hard cutoff)

Different Google sources give you a spectrum, not one magic number — and it’s directional, not a compliance guarantee. The only figure that comes from Google’s own current documentation is the 1–2 day ceiling below; the minute- and day-level figures are Google staff comments relayed by Search Engine Journal from office-hours sessions, not primary-sourced statements you can hold Google to.

• Minutes, occasionally. Gary Illyes, via Search Engine Journal: “Serving a 503 status code for an extended period of time will cause a decrease in crawl rateCrawl rate is how fast a search engine crawler fetches pages from your site — the number of simultaneous requests it makes and the delay between them. Google sets it automatically based on your server's health; it's the supply side of crawl budget, not a ranking factor.. Fortunately for you, 10-15 minutes every now and then is not ‘extended’ by any means, so you should be fine.”
• A day or so. John Mueller, via Search Engine Journal: “For an outage of maybe a day or so, using a 503 result code is a great way to tell us that we should check back.”
• 1–2 days. Google’s own documented ceiling for an urgent full-site 503 — the one figure in this list from Google’s current, primary-sourced guidance.
• “A couple of days” and beyond. This is where it turns. Mueller, via Search Engine Journal: “after a couple of days we think this is a permanent result code, and we think your pages are just gone, and we will drop them from the index.”
• Weeks. Near-guaranteed index loss, and — per the docs — recovery from a full removal has “no fixed time … and there’s no mechanism to speed that up.”

None of this adds up to a safe-harbor countdown clock. It describes what tends to happen, not a rule you can rely on for a specific outcome — treat the 1–2 day figure as the outer edge to plan around, not a guarantee that anything inside it is risk-free.

The metadata-freeze side effect (the risk nobody talks about)

Even a short, correct 503 has a cost that gets missed: while you’re serving 503s, Google can’t update what it already has. Straight from the docs: “it’s not possible for Google’s systems to refresh titles, descriptions, metadata, or structured data included on a website if a page returns a 503 HTTP response status code.”

So a 503 freezes your indexed metadata — it doesn’t erase it and it doesn’t update it. If you changed a title or a schema markupSchema markup is code that uses the schema.org vocabulary to label what your content means so search engines can understand it and show rich results. It's most often written in JSON-LD, and it's not a direct ranking factor. right before maintenance, a stale SERP snippet can persist through the whole window. “Safe” means “not permanently harmful when brief,” not “invisible.”

Never return a 503 for robots.txt

This is a real-world foot-gun: some maintenance-mode plugins and blanket CDN rules 503 everything, including robots.txt. Don’t. Google is explicit: “Don’t return a 503 HTTP response status code for the robots.txtA plain-text file at the root of a host that tells crawlers which URLs they may and may not request. It controls crawling, not indexing — a blocked URL can still be indexed if it's linked from elsewhere. file because this blocks all crawlingCrawling is how search engines use automated bots (like Googlebot and Bingbot) to discover URLs and download pages. A page has to be crawlable to be indexed, but crawling on its own isn't a ranking factor..” A 503’d robots.txt stops Google from crawlingCrawling is how search engines use automated bots (like Googlebot and Bingbot) to discover URLs and download pages. A page has to be crawlable to be indexed, but crawling on its own isn't a ranking factor. anything at all — including the recrawl that would confirm you’re back. Keep robots.txt returning 200 even during a full-site outage.

503 vs 429 — Google treats them the same

People search “429 or 503” interchangeably in rate-limiting contexts. For crawl-rate purposes they’re equivalent to Google. Mueller (on a page he himself later marked as older): “we treat them both about the same. We see both as a temporary issue, and tend to slow down crawling if we see a bunch of them.” Same backoff logic; a 429 Too Many Requests and a 503 both tell GooglebotGooglebot is Google's web crawler — the software that fetches pages so Google can index and rank them. It comes in two variants, Googlebot Smartphone (primary, under mobile-first indexing) and Googlebot Desktop, and runs an evergreen Chromium renderer. to ease off. (Its 502/504 siblings, the other transient gateway/timeout errors, get similar treatment — sustained 5xx of any flavour slows crawling.)

TIP Confirm the maintenance endpoint is really returning 503

A branded maintenance page can look correct while returning 200. Check the HTTP response itself, then confirm Retry-After separately because this result does not display the header.

Test the affected URLs with my free HTTP Status Checker Free

1. Test representative page URLs and the robots.txt URL during the maintenance window.
2. Confirm affected pages return 503 while robots.txt remains reachable, and inspect Retry-After with a header-capable client.
3. Restore 200 responses promptly and rerun the same URL list after maintenance.

Tap to enlarge

The checker shows one URL returning HTTP 503, grouped in the 5xx server-error filter. The verification panel identifies the URL and status and provides a rerun action.

Recovery is likely — not guaranteed

The most useful myth-buster in this whole topic: if a long outage does drop your pages, they usually come back, but not always the same. Mueller: “when the pages come back we will crawl them again and we will try to index them again. But it’s essentially during that time we will probably drop a lot of the pages from the website from our index, and there’s a pretty good chance that it’ll come back in a similar way but it’s not always guaranteed.” Plan your maintenance windows as if recovery could be imperfect, because it can be.

A few FAQs worth answering directly

• Does a 503 hurt SEO? Not when it’s brief and correct. A short 503 with Retry-After is the intended signal for planned downtime. The harm comes from leaving it up too long, or from 503’ing robots.txt.
• Is 503 or 404 better for maintenance? 503, every time. A 404 says “gone” and starts deindexingDeindexing means getting a URL to stop appearing in Google's search results. There's no single delete button — the right method depends on whether you own the page, whether removal is temporary or permanent, and whether the content should still exist.; a 503 says “back soon.”
• Does WordPress return a 503 during updates? Yes — WordPress serves a 503 automatically while core or plugins update. That’s the correct behaviour; the window is normally seconds.
• Can I just show a “coming soon” page at 200 instead? No. A 200 error page gets indexed as content, and if it’s the same page across every URL, Google may treat those URLs as duplicates.

The single rule to remember

Google’s own ceiling for an urgent full-site 503 is 1–2 days — treat that as the outer edge to plan around, not a guaranteed-safe window. Even inside it, a 503 still pauses your crawl rateCrawl rate is how fast a search engine crawler fetches pages from your site — the number of simultaneous requests it makes and the delay between them. Google sets it automatically based on your server's health; it's the supply side of crawl budget, not a ranking factor. and freezes your metadata, and nothing here promises a specific ranking, indexingStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed., or recovery outcome. Past that ceiling, switch to an indexable 200 placeholder before Google decides your pages are gone. Everything else on this page is a footnote to that.

For the neighbouring codes, see the rest of the HTTP status codesAn HTTP status code is the three-digit number a server returns with every response to tell a browser or crawler what happened to its request — success, redirect, client error, or server error. For SEO the code matters as much as the content: it tells Google and Bing whether to index a page, follow a redirect, retry later, or drop the URL from the index. cluster.

AI summary

A condensed take on the Advanced version:

• 503 = temporary unavailability. It’s the correct code for planned maintenance — Google prefers it over 404/410 (permanent) and over a 200 error page (which gets indexedStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed. as content, often as duplicates).
• Pair it with Retry-After (seconds or an HTTP date), though RFC 9110 makes the header optional (a MAY) — GooglebotGooglebot is Google's web crawler — the software that fetches pages so Google can index and rank them. It comes in two variants, Googlebot Smartphone (primary, under mobile-first indexing) and Googlebot Desktop, and runs an evergreen Chromium renderer. may use it as a hint for when to recrawlCrawl frequency is how often a search engine comes back to re-fetch a page it already knows about. Popular pages that change often get refreshed many times a day; stable pages can go weeks or months between crawls — and you influence it indirectly, not by setting a dial. and backs off its crawl rateCrawl rate is how fast a search engine crawler fetches pages from your site — the number of simultaneous requests it makes and the delay between them. Google sets it automatically based on your server's health; it's the supply side of crawl budget, not a ranking factor., but nothing compels an exact return time.
• Duration is a spectrum, not a guarantee: the only figure from Google’s own current docs is 1–2 days, its documented ceiling for an urgent full-site 503. Google staff comments relayed by Search Engine Journal put brief windows (10–15 minutes, ~a day) as generally fine and put “a couple of days”+ as where Google starts treating the 503 as permanent and drops pages — read those as directional, not primary-sourced commitments. For longer closures, serve an indexable 200 placeholder instead.
• Metadata freezes: while a 503 is served, Google can’t refresh titles, descriptions, or structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. — so stale snippets can persist.
• Never 503 robots.txt — that blocks all crawlingCrawling is how search engines use automated bots (like Googlebot and Bingbot) to discover URLs and download pages. A page has to be crawlable to be indexed, but crawling on its own isn't a ranking factor., including the recrawl that confirms you’re back.
• 429429 Too Many Requests is an HTTP client-error status code that means a client sent too many requests in a given time window — a mechanism called rate limiting. Unlike other 4xx codes, Google treats 429 as a server-overload signal and slows crawling instead of removing content. ≈ 503 for crawl-rate backoff; Google treats them the same.
• Recovery after a long outage is likely but “not always guaranteed” to be 1:1 (Mueller); full-index removal has no fixed recovery time.
• Best practice overall: Google’s top recommendation is to avoid a full takedown at all — limit functionality and stay live.

Official documentation

Primary-source guidance on 503 and planned downtime.

Google

• Temporarily pause or disable a website — Google’s current, canonical guidance: the preference order, the 1–2 day rule, Retry-After, the robots.txtA plain-text file at the root of a host that tells crawlers which URLs they may and may not request. It controls crawling, not indexing — a blocked URL can still be indexed if it's linked from elsewhere. exception, the metadata-freeze note, and the curl verification tip.
• How to deal with planned site downtime (2011 blog post) — the original write-up (now carries an “outdated, see best practices” banner). Source of the classic PHP Retry-After example.
• In-Depth Guide to How Google Search Works — background on crawl-rate backoff: HTTP 500A 500 Internal Server Error is a generic HTTP status code — RFC 9110 defines it as an unexpected condition that stopped the server from fulfilling the request, and nothing more. The request may have been fine — something broke server-side while generating the response. For SEO, an isolated 500 is typically retried, but persistent, site-wide 500s get Google's documented response: slower crawling and, if the errors don't clear, eventual removal from the index.-family errors mean “slow down.”

RFC / technical reference

• MDN — 503 Service Unavailable — the standards-based definition of the status code.
• MDN — Retry-After header — syntax for the seconds and HTTP-date forms.

Bing / Microsoft

• Bing does not publish 503-for-maintenance guidance at the same granularity as Google. Retry-After is a standard RFC header (not Google-specific), and BingbotBingbot is Microsoft Bing's primary web crawler — the bot that discovers, fetches, and renders pages to build the Bing index. That index also powers Yahoo, DuckDuckGo, Ecosia, and Microsoft Copilot, so Bingbot's reach is far wider than Bing's own search-market share. also backs off crawl rateCrawl rate is how fast a search engine crawler fetches pages from your site — the number of simultaneous requests it makes and the delay between them. Google sets it automatically based on your server's health; it's the supply side of crawl budget, not a ranking factor. on repeated 5xx/429 responses — but treat any specific Bing behaviour as inferred from standard HTTP semantics rather than a documented statement.

Quotes from the source

On-the-record statements from Google. Each link is a deep link that jumps to the quoted passage on the source page.

Google docs — the 1–2 day rule

• “If you need to urgently disable the site for 1-2 days, then return an informational error page with a 503 HTTP response status code instead of all content.” Jump to quote
• “If you need to disable the site for a longer time, then provide an indexable home page as a placeholder for users to find in Search by using the 200 HTTP status codeAn HTTP status code is the three-digit number a server returns with every response to tell a browser or crawler what happened to its request — success, redirect, client error, or server error. For SEO the code matters as much as the content: it tells Google and Bing whether to index a page, follow a redirect, retry later, or drop the URL from the index..” Jump to quote

Google docs — robots.txtA plain-text file at the root of a host that tells crawlers which URLs they may and may not request. It controls crawling, not indexing — a blocked URL can still be indexed if it's linked from elsewhere. and metadata

• “Don’t return a 503 HTTP response status code for the robots.txtA plain-text file at the root of a host that tells crawlers which URLs they may and may not request. It controls crawling, not indexing — a blocked URL can still be indexed if it's linked from elsewhere. file because this blocks all crawlingCrawling is how search engines use automated bots (like Googlebot and Bingbot) to discover URLs and download pages. A page has to be crawlable to be indexed, but crawling on its own isn't a ranking factor..” Jump to quote
• “it’s not possible for Google’s systems to refresh titles, descriptions, metadata, or structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. included on a website if a page returns a 503 HTTP response status code.” Jump to quote
• “removing a site completely from Google’s indexStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed. is a significant change that can take quite some time to recover from.” Jump to quote

Google Search Central Blog (2011)

• “instead of returning an HTTP result code 404 (Not FoundA 302 (\"Found\") is a temporary redirect: it forwards users to a new URL while telling search engines the original URL should stay in the index. It's a weak canonicalization signal, not the zero-equity dead end of SEO folklore.) or showing an error page with the status code 200 (OK) when a page is requested, it’s better to return a 503 HTTP result code (Service Unavailable) which tells search engine crawlersA crawler — also called a spider or bot — is an automated program that fetches web pages, extracts their links, and queues new URLs to visit. Search engines use crawlers to discover and download content for their index. that the downtime is temporary.” Jump to quote
• “If known, the length of the downtime in seconds or the estimated date and time when the downtime will be complete can be specified in an optional Retry-After header, which GooglebotGooglebot is Google's web crawler — the software that fetches pages so Google can index and rank them. It comes in two variants, Googlebot Smartphone (primary, under mobile-first indexing) and Googlebot Desktop, and runs an evergreen Chromium renderer. may use to determine when to recrawlCrawl frequency is how often a search engine comes back to re-fetch a page it already knows about. Popular pages that change often get refreshed many times a day; stable pages can go weeks or months between crawls — and you influence it indirectly, not by setting a dial. the URL.” Jump to quote

John Mueller, Google

• “please return a ‘503 Service unavailable503 Service Unavailable is the HTTP status code a server returns when it's temporarily unable to handle a request — usually because of overload or planned maintenance. It tells crawlers 'come back later' instead of 'this page is gone,' which is why Google recommends it (paired with a Retry-After header) for short, planned downtime.’ HTTP result code… they’re generally more than happy to give your site some time to catch up again.” Jump to quote From johnmu.com/503s/ (2013), Mueller’s personal republish of a Google+ post. The live page renders typographic apostrophes, so the deep link is kept short to guarantee an exact text match.
• “we treat them both about the same. We see both as a temporary issue, and tend to slow down crawlingCrawling is how search engines use automated bots (like Googlebot and Bingbot) to discover URLs and download pages. A page has to be crawlable to be indexed, but crawling on its own isn't a ranking factor. if we see a bunch of them.” (on 429 vs 503) Jump to quote From johnmu.com/429-or-503/ (2015), a page Mueller himself marks as “old, probably outdated.”
• “after a couple of days we think this is a permanent result code, and we think your pages are just gone, and we will drop them from the index.” Jump to quote Relayed by Search Engine Journal from a Search Central office-hours/Q&A; confirm against the original before treating as final.

Gary Illyes, Google

• “Serving a 503 status code for an extended period of time will cause a decrease in crawl rateCrawl rate is how fast a search engine crawler fetches pages from your site — the number of simultaneous requests it makes and the delay between them. Google sets it automatically based on your server's health; it's the supply side of crawl budget, not a ranking factor.. Fortunately for you, 10-15 minutes every now and then is not ‘extended’ by any means, so you should be fine.” Jump to quote Relayed by Search Engine Journal from a Google Search Central SEO office-hours session.

Planned-maintenance 503 checklist

Run this before, during, and after a maintenance window:

• Decided whether you even need a full takedown — can you limit functionality and stay live instead (Google’s preferred option)?
• Maintenance page returns a real 503 status (verified with curl, not just eyeballed in a browser).
• A Retry-After header is set (seconds or an HTTP date) with a realistic estimate.
• robots.txt still returns 200 and is reachable — never 503’d.
• The 503 page is lightweight: static HTML, inline CSS, no heavy assets (botsA crawler — also called a spider or bot — is an automated program that fetches web pages, extracts their links, and queues new URLs to visit. Search engines use crawlers to discover and download content for their index. will hit it repeatedly).
• A hard internal deadline is set to switch to a 200 indexable placeholder if the window exceeds ~1–2 days.
• No pages are returning 404/410 or a 200 “coming soon” page instead of the 503.
• Search ConsoleA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. Page IndexingThe Google Search Console report (formerly Index Coverage) showing how many of your URLs are indexed vs. not indexed, and grouping the not-indexed ones by reason. and Crawl StatsA Google Search Console report (under Settings) that shows how Google has crawled your site over the last 90 days — total requests, download size, and average response time, broken down by response code, file type, Googlebot type, and purpose. It's only available for root-level properties (a Domain property or a URL-prefix property verified at the site's root). are being watched during and after the window for 5xx spikes.
• Post-maintenance: confirmed pages returned to 200, submitted a fresh crawl via URL InspectionA Google Search Console feature that reports how Google sees one specific URL on a property you own. By default it shows the last-indexed snapshot; a separate \"Test live URL\" mode fetches the current version. for key URLs, and verified metadata/snippets refreshed.

SOP: Put your site into planned maintenance mode correctly

A repeatable procedure for a short, safe maintenance window.

Before the window

1. Decide the scope. If you can keep the site live and only disable the risky part (cart, checkout, a broken feature), do that instead — it’s Google’s top recommendation and avoids all of the risk below. Only proceed to a full 503 if you genuinely must take everything offline.
2. Estimate the duration honestly. If it’s realistically more than 1–2 days, don’t plan a full-site 503 at all — plan an indexable 200 placeholder homepage instead.
3. Build a lightweight maintenance page. Static HTML, inline CSS, base64 or no images. BotsA crawler — also called a spider or bot — is an automated program that fetches web pages, extracts their links, and queues new URLs to visit. Search engines use crawlers to discover and download content for their index. (and users) will hit it repeatedly, so it must not strain the server.

Configuring the response

4. Return a real 503 for site content. Examples:
   • Apache (.htaccess): route everything to the maintenance page and force the status —

     RewriteEngine On
     RewriteCond %{REQUEST_URI} !^/maintenance\.html$
     RewriteCond %{REQUEST_URI} !^/robots\.txt$
     RewriteRule ^ /maintenance.html [R=503,L]
     ErrorDocument 503 /maintenance.html
     Header always set Retry-After "3600"

   • Nginx:

     location / {
         return 503;
     }
     error_page 503 /maintenance.html;
     location = /maintenance.html { internal; add_header Retry-After 3600; }
     location = /robots.txt { }   # keep robots.txt serving 200

   • PHP (Google’s own example form):

     header('HTTP/1.1 503 Service Unavailable');
     header('Retry-After: 3600');

5. Set Retry-After to a realistic estimate — seconds (3600) or an HTTP date.
6. Exempt robots.txt. Make sure robots.txt still returns 200 (see the rewrite conditions above). This is the step most people miss.

Verify

7. Confirm the status with curl, not the browser:

   curl -I -X GET "https://www.example.com/"
   # expect: HTTP/1.1 503 Service Unavailable  +  Retry-After: ...
   curl -I "https://www.example.com/robots.txt"
   # expect: HTTP/1.1 200 OK

During and after

8. Watch the clock. If you’re approaching the 1–2 day ceiling, switch to a 200 indexable placeholder before Google starts treating the 503 as permanent.
9. Bring the site back. Remove the 503 rule; confirm pages return 200 again.
10. Recheck. Use GSCA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. URL InspectionA Google Search Console feature that reports how Google sees one specific URL on a property you own. By default it shows the last-indexed snapshot; a separate \"Test live URL\" mode fetches the current version. on key pages, watch Crawl StatsA Google Search Console report (under Settings) that shows how Google has crawled your site over the last 90 days — total requests, download size, and average response time, broken down by response code, file type, Googlebot type, and purpose. It's only available for root-level properties (a Domain property or a URL-prefix property verified at the site's root). for the 5xx spike to subside, and confirm titles/descriptions/structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. refresh (they were frozen during the 503).

What not to do — and the myths behind it

Myth: “A 503 for more than a couple of days is completely safe as long as I use Retry-After.” Retry-After improves precision; it doesn’t override Google’s ceiling. The docs cap an urgent full-site 503 at 1–2 days, and Mueller: “after a couple of days we think this is a permanent result code… and we will drop them from the indexStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed..” Switch to a 200 placeholder before you get there.

Myth: “503 has zero SEO impact if implemented correctly.” Even a short, correct 503 pauses metadata refresh — titles, descriptions, and structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. can’t update while it’s served — and any 503 nudges crawl rateCrawl rate is how fast a search engine crawler fetches pages from your site — the number of simultaneous requests it makes and the delay between them. Google sets it automatically based on your server's health; it's the supply side of crawl budget, not a ranking factor. down. “Safe” means “not permanently harmful when brief,” not “invisible.”

Myth: “A 404 or a 200 ‘coming soon’ page is just as good during maintenance.” No. 404/410 signals permanent removal and gets pages dropped faster. A blanket 200 error page gets indexed as content — and if every URL returns the same page, Google may treat them as duplicates.

Myth: “You should 503 your robots.txt too, to be thorough.” The opposite. A 503’d robots.txt blocks all crawlingCrawling is how search engines use automated bots (like Googlebot and Bingbot) to discover URLs and download pages. A page has to be crawlable to be indexed, but crawling on its own isn't a ranking factor., including the recrawlCrawl frequency is how often a search engine comes back to re-fetch a page it already knows about. Popular pages that change often get refreshed many times a day; stable pages can go weeks or months between crawls — and you influence it indirectly, not by setting a dial. that confirms your site is back. Keep it at 200.

Myth: “If Google drops my pages during downtime, they’ll come back exactly the same.” Likely, but “not always guaranteed” (Mueller), and full-removal recovery has no fixed timeline and no way to speed it up.

Myth: “429429 Too Many Requests is an HTTP client-error status code that means a client sent too many requests in a given time window — a mechanism called rate limiting. Unlike other 4xx codes, Google treats 429 as a server-overload signal and slows crawling instead of removing content. and 503 are totally different signals.” For crawl-rate backoff, Google “treat[s] them both about the same.”

The anti-pattern to burn into memory: serving a bare 200 OK “down for maintenance” page. It’s the single most common mistake and the one Google warns about most directly.

503 for maintenance — cheat sheet

Which code for which situation

Duration spectrum

The 1–2 day row is Google’s own current documented ceiling. The minute- and day-level rows are Google staff comments relayed by Search Engine Journal, not primary Google documentation — read the spectrum as directional, not a guaranteed-safe compliance window.

Retry-After forms

• Seconds: Retry-After: 3600
• HTTP date: Retry-After: Sat, 8 Oct 2011 18:27:00 GMT

Don’t-forget rules

• robots.txt stays 200 — never 503 it.
• 503 page = static HTML, inline CSS (botsA crawler — also called a spider or bot — is an automated program that fetches web pages, extracts their links, and queues new URLs to visit. Search engines use crawlers to discover and download content for their index. hit it repeatedly).
• 503 freezes indexed titles/descriptions/schema — they won’t refresh.
• Verify with curl -I, not the browser.

Verify your 503 is actually a 503

A browser can show a maintenance page while the server quietly returns 200. Check the real status code and headers with curl.

macOS / Linux

# Site content should return 503 with a Retry-After header
curl -I -X GET "https://www.example.com/"
# expect: HTTP/1.1 503 Service Unavailable
#         Retry-After: 3600

# robots.txt must stay reachable (200), NOT 503
curl -I "https://www.example.com/robots.txt"
# expect: HTTP/1.1 200 OK

Windows (PowerShell)

# -SkipHttpErrorCheck lets PowerShell show the 503 instead of throwing
(Invoke-WebRequest -Uri "https://www.example.com/" -Method Head -SkipHttpErrorCheck).StatusCode
(Invoke-WebRequest -Uri "https://www.example.com/robots.txt" -Method Head).StatusCode

Minimal 503 responses

PHP (Google’s own example form)

<?php
header('HTTP/1.1 503 Service Unavailable');
header('Retry-After: 3600'); // or an HTTP date: 'Sat, 8 Oct 2011 18:27:00 GMT'
?>
<!DOCTYPE html>
<title>We'll be right back</title>
<h1>Down for scheduled maintenance</h1>
<p>We expect to be back within the hour. Thanks for your patience.</p>

Nginx (content 503, robots.txtA plain-text file at the root of a host that tells crawlers which URLs they may and may not request. It controls crawling, not indexing — a blocked URL can still be indexed if it's linked from elsewhere. exempt)

location / {
    return 503;
}
error_page 503 /maintenance.html;
location = /maintenance.html {
    internal;
    add_header Retry-After 3600;
}
location = /robots.txt { }   # keep serving robots.txt normally (200)

Keep the maintenance page static with inline CSS and no heavy assets — botsA crawler — also called a spider or bot — is an automated program that fetches web pages, extracts their links, and queues new URLs to visit. Search engines use crawlers to discover and download content for their index. will request it repeatedly and you don’t want that page itself straining the server.

Tools for handling and checking a 503

• curl -I — the fastest way to confirm the real status code and Retry-After header (a browser can hide the truth). See the Scripts tab.
• Google Search ConsoleA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. — URL InspectionA Google Search Console feature that reports how Google sees one specific URL on a property you own. By default it shows the last-indexed snapshot; a separate \"Test live URL\" mode fetches the current version. — check the last crawled status of a specific URL and request a fresh crawl once you’re back online.
• GSC — Crawl Stats reportA Google Search Console report (under Settings) that shows how Google has crawled your site over the last 90 days — total requests, download size, and average response time, broken down by response code, file type, Googlebot type, and purpose. It's only available for root-level properties (a Domain property or a URL-prefix property verified at the site's root). — watch for the 5xx/503 spike during the window and confirm it subsides afterward.
• GSC — Page Indexing reportThe Google Search Console report (formerly Index Coverage) showing how many of your URLs are indexed vs. not indexed, and grouping the not-indexed ones by reason. — spot any pages that got dropped if the window ran long.
• Screaming Frog SEO SpiderA crawler — also called a spider or bot — is an automated program that fetches web pages, extracts their links, and queues new URLs to visit. Search engines use crawlers to discover and download content for their index. — crawl your own site to confirm which URLs return 503 vs 200 (and that robots.txt isn’t 503’d).
• Ahrefs Site Audit / Ahrefs Webmaster ToolsThe free, first-party consoles search engines give site owners — Google Search Console and Bing Webmaster Tools — to see how their pages are crawled, indexed, and ranked, and to fix problems. The legacy name is 'webmaster tools.' — surfaces 5xx responses across the site so you can catch a stray 503 that lingers past the window.
• WordPress maintenance-mode plugins (e.g. WP Maintenance Mode) — WordPress already 503s automatically during core/plugin updates; a plugin gives you a controlled maintenance page — just confirm it doesn’t 503 robots.txt.

Resources worth your time

My related writing

• HTTP Status Codes: The Complete List for SEO — my full guide to the status codes that matter for SEO, with 503 in the 5xx context. This single-topic deep dive is the expanded version of that guide’s one-line treatment.
• The Beginner’s Guide to Technical SEO — where crawlingCrawling is how search engines use automated bots (like Googlebot and Bingbot) to discover URLs and download pages. A page has to be crawlable to be indexed, but crawling on its own isn't a ranking factor., status codes, and maintenance fit in the bigger picture.

From around the industry

• Temporarily pause or disable a website (Google Search Central) — the canonical, current guidance; read this first.
• How to deal with planned site downtime (Google Search Central Blog, 2011) — the original write-up and the classic PHP Retry-After example.
• 503s (John Mueller, johnmu.com) — Mueller’s own plain-language take on why to 503 during downtime.
• Google On The SEO Impact Of 503 Status Codes (Search Engine Journal) — Gary Illyes on the “10–15 minutes is fine” data point.
• Google Will Deindex Pages If Site Is Down For Several Days (Search Engine Journal) — Mueller’s “a couple of days… we will drop them” threshold.
• HTTP 503: Handling site maintenance correctly for SEO (Yoast) — a practical, WordPress-flavoured walkthrough of the Retry-After mechanics.
• 503 Service Unavailable (MDN) — the standards reference for the status code itself.

Videos

• Google Search Central (YouTube) — the How Google Search WorksSearch works in three stages — crawling, indexing, and serving (ranking). A page has to clear each one to appear in results: getting crawled doesn't mean you're indexed, and getting indexed doesn't mean you rank. series and the SEO office-hours archives, where Mueller and Illyes have answered the 503/maintenance and deindexingDeindexing means getting a URL to stop appearing in Google's search results. There's no single delete button — the right method depends on whether you own the page, whether removal is temporary or permanent, and whether the content should still exist.-timeline questions quoted throughout this page. Channel

Stats worth citing

• 1–2 days — Google’s documented ceiling for using a full-site 503 during an urgent takedown; beyond that, switch to a 200 placeholder. Source
• 10–15 minutes, occasionally = fine — Gary Illyes on how brief a 503 window can be without denting crawl rateCrawl rate is how fast a search engine crawler fetches pages from your site — the number of simultaneous requests it makes and the delay between them. Google sets it automatically based on your server's health; it's the supply side of crawl budget, not a ranking factor.. Source
• “A couple of days” — the point at which Google starts treating a 503 as permanent and dropping pages from the indexStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed. (John Mueller). Source
• No fixed recovery time — Google’s docs on recovering from a complete removal from the index: no set timeline and no way to speed it up. Source

Test yourself: 503 Service Unavailable

Five quick questions on using 503 for planned maintenance. Pick an answer for each, then check.

503 Service Unavailable

503 Service Unavailable is the HTTP status code a server returns when it's temporarily unable to handle a request — usually because of overload or planned maintenance. It tells crawlers 'come back later' instead of 'this page is gone,' which is why Google recommends it (paired with a Retry-After header) for short, planned downtime.

Related: HTTP Status Code, 500 Internal Server Error, 429 Too Many Requests

503 Service Unavailable

503 Service Unavailable is a 5xx HTTP status codeAn HTTP status code is the three-digit number a server returns with every response to tell a browser or crawler what happened to its request — success, redirect, client error, or server error. For SEO the code matters as much as the content: it tells Google and Bing whether to index a page, follow a redirect, retry later, or drop the URL from the index. that means the server understood the request but can’t handle it right now because it’s temporarily overloaded or down for maintenance. The “temporarily” is the whole point: unlike a 404/410 (permanent removal) or a 200 with error content (which reads as real content), a 503 tells search engine crawlersA crawler — also called a spider or bot — is an automated program that fetches web pages, extracts their links, and queues new URLs to visit. Search engines use crawlers to discover and download content for their index. the outage is short-lived and they should check back later.

Because of that semantic, Google explicitly recommends returning a 503 — not a 404 and not a 200 “coming soon” page — when you have to take a site down for planned maintenance. Google’s own current guidance sets 1–2 days as the ceiling for an urgent full-site 503 — treat that as the outer edge to plan around, not a guaranteed-safe window; serve one for much longer and Google starts to read it as permanent and can drop your pages from the indexStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed..

Two implementation details matter. First, pair the 503 with a Retry-After header (a number of seconds or an HTTP date) so GooglebotGooglebot is Google's web crawler — the software that fetches pages so Google can index and rank them. It comes in two variants, Googlebot Smartphone (primary, under mobile-first indexing) and Googlebot Desktop, and runs an evergreen Chromium renderer. has a hint about when to return and can back its crawl rateCrawl rate is how fast a search engine crawler fetches pages from your site — the number of simultaneous requests it makes and the delay between them. Google sets it automatically based on your server's health; it's the supply side of crawl budget, not a ranking factor. off accordingly. Second, never return a 503 for robots.txt itself — a 503’d robots.txtA plain-text file at the root of a host that tells crawlers which URLs they may and may not request. It controls crawling, not indexing — a blocked URL can still be indexed if it's linked from elsewhere. blocks all crawlingCrawling is how search engines use automated bots (like Googlebot and Bingbot) to discover URLs and download pages. A page has to be crawlable to be indexed, but crawling on its own isn't a ranking factor., including the recrawlCrawl frequency is how often a search engine comes back to re-fetch a page it already knows about. Popular pages that change often get refreshed many times a day; stable pages can go weeks or months between crawls — and you influence it indirectly, not by setting a dial. that would confirm your site is back. Google treats 503 and 429 Too Many Requests about the same way for crawl-rate purposes; both are “slow down” signals, alongside its 500-family siblings like 502 and 504.

Related: HTTP Status Code, 500 Internal Server Error, 429 Too Many Requests

Changelog