November has been a rough month for the software that runs the web. On November 11, Apache fixed a critical flaw in its HTTP/2 module. A week later, nginx disclosed five more, quickly nicknamed "NGINX2Root." Together, Apache and nginx serve a huge share of the websites on the internet, very likely including yours.

If you own a WordPress site, you can't patch either one yourself, and most people have no idea who can. That's the part I want to clear up, because the useful question isn't "are we covered?" It's "what version are you running, and when did it go on?"

The short version: the November 2026 nginx vulnerability (NGINX2Root, led by CVE-2026-123007) is fixed in nginx 1.30.7 and 1.31.9. The Apache vulnerability (CVE-2026-119618) is fixed in Apache 2.4.71 and only affects servers with Apache's HTTP/2 module on. Neither is a WordPress bug, and neither can be fixed from your WordPress dashboard. On shared or managed hosting, your host patches the web server; your job is to ask which version you're on and when it was installed.

What actually happened in November

Apache, CVE-2026-119618. A flaw in Apache's HTTP/2 module (mod_http2) can let an attacker run code on the server without logging in. Apache rates it critical. It affects versions 2.4.55 through 2.4.70 and was fixed in 2.4.71 on November 11. On November 17, Singapore's national cyber agency reported that it was being exploited and that proof-of-concept code was public. Only servers with Apache's HTTP/2 module switched on are exposed.

nginx, "NGINX2Root." On November 18, nginx and F5 published fixes for five flaws, led by CVE-2026-123007, a memory corruption bug in nginx's HTTP/2 handling. The fixed versions are 1.30.7 (stable) and 1.31.9 (mainline). Censys counted about 8.8 million internet-facing services on potentially affected versions with HTTP/2 on. As of November 19, nobody had confirmed attacks in the wild, and fake exploit code was already circulating on GitHub, which tells you how fast attention moves now.

nginx, "Rift." This one came earlier, in May: CVE-2026-42945, a buffer overflow in nginx's rewrite module, fixed in 1.30.1 and 1.31.0. Exploit attempts showed up on security researchers' decoy servers within days of disclosure.

A detail worth noticing: the Apache bug was credited to Anthropic's automated scanner and to Google's vulnerability researchers, who found it independently. Rift was found by an AI-driven detection platform. I'll come back to why that matters.

Why a WordPress owner should care

WordPress doesn't run on its own. Every request to your site passes through a web server first, usually nginx, Apache, or both, before WordPress ever sees it. A flaw at that layer sits underneath every site on the machine. Your plugins can be perfectly up to date and it won't matter.

Most of the security advice written for WordPress owners is about WordPress: updates, plugins, passwords, 2FA. All of that matters, and we write about it a lot. But a web server flaw is a different problem with a different owner.

A website request passing from a laptop through a web server before reaching the WordPress site

Who patches what

This is where most people get confused, so here it is plainly:

  • The server layer (the operating system, the web server, PHP) belongs to whoever runs the server. On managed hosting, that's your host. On a VPS or a "you manage it" cloud server, that's you or your developer, whether anyone told you or not.
  • The application layer (WordPress core, plugins, themes) belongs to whoever manages the site. With us, that's us. With most hosts, it's you.

A November-style web server flaw lands entirely in the first bucket. If you're on shared or managed hosting, there's nothing for you to click. Your only job is to find out whether your host did theirs. (If you'd rather have both layers handled by one team, that's how our managed WordPress security works.)

A website owner at a desk and a hosting engineer at a server, showing who is responsible for which layer

What we saw on our own servers

Here's our timeline, from the servers' own package logs. Our hosting platform delivers these patches through its managed server packages. Our job is to confirm they actually landed on every server, and to chase down the ones that didn't.

Apache 2.4.71. Reached our servers between November 13 and November 17, two to six days after the fix shipped. The last one updated on November 17, the same day exploitation was publicly reported. There's a second layer here worth explaining: on our platform, nginx faces the internet and Apache sits behind it. On the servers we checked, Apache's HTTP/2 module isn't even switched on, which means the vulnerable code most likely wasn't reachable in the first place. Architecture is a security control too.

NGINX2Root. Patched across all of our servers within about a day of disclosure. On most of them that meant nginx 1.31.9 by November 19; on one older server due to move to new hardware, the fix was switching off HTTP/2, where the most serious of these bugs live.

Rift. When we went back and checked months later, every server was on a fixed version. I can't give you the exact day it landed in May, because the logs don't go back that far.

We've been through this recently with PHP. In October, a critical flaw in PHP itself (the language WordPress is written in) went public. The fixed PHP was running on our servers about 10 to 17 hours after the public release. We got ten tickets from ten clients in four days asking if they were OK, and every one got an answer the same day. Around the same time, a nonprofit came to us on a sales call because their previous host had taken a week to patch the same flaw. A week is a long time when exploit code shows up within days.

Does Cloudflare protect you?

Partly. If your site is behind Cloudflare's proxy, as every site we host is, visitors' HTTP/2 connections end at Cloudflare's edge, not at your server. Cloudflare's own write-up says that protects origins from the HTTP/2 and HTTP/3 bugs in NGINX2Root.

The catch is in the fine print: that only holds if nobody can reach your server directly. Servers often have addresses that skip the CDN entirely, and attackers know how to find them. A CDN in front of your site is a strong layer. It isn't a reason to leave the server behind it unpatched.

Website traffic passing through an edge network before reaching the origin server

My take: hosts should publish this

When something this big drops, your host should tell you what they did and when. Most don't. Of the big managed WordPress hosts I checked, I found exactly one, SiteGround, that published a statement on NGINX2Root, saying it patched 30 minutes after release. I can't verify their number, but I respect that they put one out. Silence leaves you guessing, and guessing isn't a security plan.

I'll also say what I expect, because I think it's a fair professional call: we're going to see more of these, not fewer. More than 100,000 vulnerabilities have been published in 2026, more than double last year, and AI tools are now finding real bugs in software as mature as Apache. Only a small share of all those bugs ever get exploited. But the ones that do get exploited faster, and the job for every host is now speed: find out, patch, verify, tell people. I hope I'm wrong about the volume. I don't think I am.

Three questions to ask your host this week

  1. "Which Apache and nginx versions are my sites running right now, and when did they go on?" A good host answers with versions and dates, not "we're on top of it."
  2. "Is HTTP/2 handled by nginx, Apache, or both, and is my server reachable without going through the CDN?" This tells you how much the CDN is really protecting you.
  3. "When something like this happens, how will you tell me?" If the answer is "you'd have to ask," you just learned something.

If you're on a server you manage yourself, the questions are the same. You're just asking them of yourself, or of whoever set the server up.

For how this fits with plugins, firewalls and everything else, see how the layers of WordPress security fit together and the threats WordPress sites actually face. If you'd rather not track any of this yourself, that's what our managed WordPress security is for.

FAQ

Is my WordPress site affected by the nginx vulnerability?

Only if your server runs an affected nginx version (1.9.5 through 1.31.8 for the lead flaw, CVE-2026-123007) and your host hasn't updated it to 1.30.7 or 1.31.9 or later. WordPress itself isn't the problem, and there's nothing to fix in your WordPress dashboard. Ask your host which version you're on.

Is my WordPress site affected by the Apache vulnerability?

Only if your server runs Apache 2.4.55 through 2.4.70 with the HTTP/2 module enabled. Many managed hosts put nginx in front of Apache, and on those setups Apache's HTTP/2 module is often off. The fix is Apache 2.4.71, and it's your host's job to install it.

Can I patch Apache or nginx myself on shared or managed hosting?

No. On shared or managed hosting, the web server belongs to the host. You can't update it, and you shouldn't need to. You can ask your host for the version and the date it was updated.

Does Cloudflare protect my site from NGINX2Root?

Cloudflare's proxy protects against the HTTP/2 and HTTP/3 flaws, because those connections end at Cloudflare instead of your server. That protection only holds if your server can't be reached directly, so it's a layer, not a substitute for patching.

How fast should a host patch a critical web server vulnerability?

Within hours to a few days of a fix being available, and they should be able to show you when. Exploit attempts against these flaws have shown up within days of disclosure, so a week is too slow.