Google’s PHP Patch Advice Is Really a Crawl Budget Argument

Google is telling site owners to return 503 status codes while they take a site down to apply PHP patches. The guidance surfaced in Search Engine Journal’s SEO Pulse roundup, which also covered robots.txt controls, how Google treats bracket URLs, and JPEG XL support in Google Images and Discover.

What Google actually said

The 503 recommendation is the headline item, and it’s narrower than it sounds. Google’s position is that a site going offline temporarily to patch PHP should signal that with a 503 response rather than letting requests fail in other ways. The roundup frames it as a reminder, which is telling: this is not new policy, it’s Google restating something that sites keep getting wrong.

The other three items in the same roundup are less dramatic but arguably more useful day to day. Google addressed the limits of what robots.txt can control, how it handles URLs containing brackets, and confirmed JPEG XL support across Google Images and Discover. Each of those touches a different part of the stack, and none of them are the kind of thing that gets a site penalized overnight. They’re the kind of thing that quietly costs you impressions for a year before anyone notices.

Why the 503 advice matters more than it looks

Here’s the part worth sitting with. A 503 with a Retry-After header tells Googlebot the site is temporarily unavailable and to come back later. A 500, a 502, a connection reset, or a firewall page that returns 200 with an error message tells Googlebot something else entirely, and none of those messages are “come back tomorrow.”

If you’re patching PHP across a fleet of sites, the difference between those responses is the difference between a crawl pause and a crawl problem. Sustained server errors can push Google toward slowing crawl rates, and in bad cases toward treating pages as unavailable. A clean 503 is the polite version of “we’re closed, come back at 3pm.” It’s a small implementation detail with outsized consequences, and it’s the sort of thing that gets skipped when a patch window is scheduled for 2am and everyone just wants it done.

My read: the fact that Google is restating this now suggests enough sites are still getting it wrong to be worth a public reminder. If your deploy process takes the origin offline without a maintenance response, that’s a gap worth closing before the next patch cycle, not after.

Robots.txt limits and bracket URLs

The robots.txt item is the one I’d watch most closely, because it touches a persistent misconception. Robots.txt controls crawling, not indexing. A page blocked in robots.txt can still appear in search results if it’s linked from elsewhere, and the snippet Google shows may come from that external context rather than the page itself. Site owners who reach for robots.txt to keep something out of the index are often solving the wrong problem, and Google’s guidance on the limits of those controls is a reminder that the two systems are separate.

Bracket URLs are a smaller curiosity but a real one. URLs containing square brackets show up in faceted navigation, in some CMS-generated paths, and in legacy structures. Google’s explanation of how it treats them matters if any of your URLs carry that punctuation, because the answer determines whether those pages get crawled and consolidated the way you expect or get handled on a different track.

JPEG XL in Images and Discover

JPEG XL support in Google Images and Discover is the most concrete item in the roundup and the easiest to act on. If Google’s image surfaces can handle the format, sites carrying large image libraries have a real compression option that wasn’t previously worth considering for search-visible assets. That’s a performance and bandwidth question as much as an SEO one, and it’s the kind of change that pays off slowly across a big catalog.

I wouldn’t rush a format migration on the strength of one roundup item. But if you’re already re-encoding images for Core Web Vitals work, JPEG XL just moved from “maybe later” to “worth testing.”

What to watch next

None of these four items are urgent in isolation. Together they sketch a familiar pattern: Google publishing clarifications on mechanics that site owners routinely mishandle, in a format that’s easy to skim and easy to ignore. The 503 guidance is the one with real downside if you get it wrong, because it’s tied to an operational moment, a patch window, when nobody is thinking about crawl behavior.

The practical move is to check what your origin actually returns during maintenance. Not what you assume it returns. If it’s anything other than a 503 with a retry hint, fix that before the next PHP release lands, because the next roundup on this topic will probably say the same thing again.