NGINX CVE

de2

Customer
Joined
Feb 9, 2010
Messages
3
Reaction score
1
Dear 3CX Security Team,


Following the disclosure of multiple critical NGINX vulnerabilities on 18 November 2026, we would like to know whether 3CX installations are affected.


We operate a self-hosted 3CX Enterprise V20 installation on Debian 12 in Microsoft Azure, deployed using the official 3CX image.


As 3CX now maintains its own embedded NGINX version, could you please clarify:


  1. Is the NGINX version bundled with 3CX affected by these vulnerabilities?
  2. If so, when can we expect a security hotfix?
  3. Are any temporary mitigation measures recommended?
  4. Will the security fix be deployed automatically?

Our PBX is internet-facing, so we would appreciate a timely response.


Kind regards,
 
  • Like
Reactions: Charli
For the reported CVE's nginx has released an update however this is not yet released from Debian, we are currently checking also on our side and get back with an update as soon as possible.
 
Thank you for your response.


However, according to your own security announcement from June 2026, 3CX decided to bundle and maintain its own customized NGINX version, specifically to avoid depending on Debian’s security update schedule.


https://www.3cx.com/blog/news/http2-vulnerability-mitigation/


You also stated that this approach would allow 3CX to respond immediately to future NGINX vulnerabilities.
 
Yes of course as blog post says which give us the ability to make updates without waiting for Debian.
In any case though we need to evaluate fixes before we push any update and not just blindly send an update.
We currently checking and we will inform once there is an update.
 
  • Like
Reactions: Bazzie79
Hi,
Bumping this thread – we are in the same situation and manage several internet-facing 3CX PBXs.

For reference, CERT Polska has published an advisory on these issues ("nginx2root"):
https://moje.cert.pl/komunikaty/202...ozliwiajace-zdalne-wykonanie-kodu-nginx2root/

It covers five CVEs, all fixed in nginx 1.31.9:
- CVE-2026-123007 – HTTP/2 (nginx 1.9.5 – 1.31.8)
- CVE-2026-123633 – HTTP/2 + proxying to an upstream server (from 1.19.7)
- CVE-2026-123634 – resolver + proxy_pass + mirror / subrequests (from 1.13.4)
- CVE-2026-121810 – HTTP/3 (QUIC) + proxying to an upstream server (from 1.25.0)
- CVE-2026-113425 – mail proxy with auth_http (from 0.5.15)

Could 3CX please give a clear statement on the following:

1. Affected versions
Which 3CX versions/builds (V20 Update x, V18 if still relevant; Debian and Windows; self-hosted and 3CX-hosted) ship a vulnerable nginx build, and which of the five CVEs actually apply to the nginx configuration used by 3CX?

2. Fix
Which 3CX versions will receive the patched nginx, and what is the expected ETA? Will it be installed automatically (with the regular 3CX/OS updates), or do we have to trigger it manually? Does it require a service restart or cause downtime?

3. Communication
Where and how will you announce the fix (blog, security advisory page, this thread, e-mail to partners, notification in the Management Console)? How can we verify that a given PBX already runs the patched nginx (package/version number)?

4. Interim mitigation, until the fix is available
- Is disabling HTTP/2 in the nginx configuration supported by 3CX and safe? Will it break anything – Web Client, iOS/Android apps, WebRTC, phone provisioning, SBC/Bridge connections, CRM/Teams integrations, or anything relying on gRPC (CERT notes that gRPC stops working with HTTP/2 disabled)?
- Is disabling HTTP/2 sufficient for 3CX? According to CERT it only mitigates CVE-2026-123007 and CVE-2026-123633. Does the 3CX configuration use HTTP/3/QUIC, "resolver" with "proxy_pass" to hostnames, "mirror"/SSI, or the mail proxy module? If so, what additional steps do you recommend?
- Will manual changes to the nginx configuration be overwritten by 3CX updates or by changes made in the Management Console? Where exactly should the change be made so that it persists?
- Do you recommend any other temporary measures (e.g. restricting access to the HTTPS port by source IP)?

Given the severity (remote code execution on an internet-facing PBX), an official advisory with a timeline would be much appreciated – even if for now it only says "under evaluation, next update on <date>".

Thank you.
 
As a temporary security measure, we have restricted access to TCP port 443 to trusted public IP addresses for one of our customers.


We tested the 3CX mobile app from an external, non-whitelisted IP address.


Results:


● Incoming and outgoing calls: Working.


● Push notifications: Working.


● Queue login using *62: Working.


● Queue logout using *63: Working.


● Presence information: Not available.


● Queue button in the app: Not working.


For our customer, this is an acceptable temporary workaround. Essential telephony functions remain operational while public exposure is significantly reduced.


We will maintain this restriction until 3CX releases a security fix.
 

Forum statistics

Threads
112,549
Messages
593,175
Members
165,760
Latest member
malo