Security AlertsThu, 19 Nov 2026
NGINX HTTP/2 vulnerabilities: update web, API and embedded proxy deployments
Two HTTP/2 flaws can cause memory corruption and remote code execution under specific conditions. Inventory exposed endpoints and dependencies.
Article
On 18 November 2026, F5 disclosed NGINX HTTP/2 security fixes. CVE-2026-123007 is a flow-control memory corruption flaw; CVE-2026-123633 is a connection use-after-free. Both require HTTP/2 enabled. Successful exploitation can run code with NGINX worker privileges, disclose data or disrupt service. The F5 CNA rates each CVSS v3.1 8.1 and CVSS v4.0 9.2.
For NGINX Open Source, CVE-2026-123007 affects 1.9.5 through 1.31.8, and CVE-2026-123633 affects 1.19.7 through 1.31.8. Fixed releases are 1.30.7 stable and 1.31.9 mainline. NGINX Plus releases fixing both include 37.0.8.1 and 37.3.0.1. Plan migration of older R33 to R36 releases according to F5 support guidance. Distribution packages can contain backports; check the distributor advisory.
Rapid7 demonstrated CVE-2026-123007 RCE in an NGINX 1.31.8 container with HTTP/2 and TLS enabled. Its method needs roughly 128 GB of traffic, is constrained by bandwidth and connection timeouts, and relies on a specific x86-64 glibc memcpy implementation and CPU features. These constraints describe that demonstration, not every possible exploit. CVE-2026-123633 involves reverse-proxy processing and scarce connection resources. A research demonstration is not proof of exploitation in the wild.
- 01
Reach affected entry
- 02
Trigger memory flaw
- 03
Execute as worker
- 04
Impact services and data
Action: investigate from the external edge to internal assets
- 01
Inventory exposure and dependencies
Inventory public and internal NGINX, outsourced systems, container images, ingress, reverse proxies and embedded versions; check HTTP/2 and vendor patches.
See what to doResources: 1
TWISA member solutions
Inventory exposure and dependencies
Inventory public and internal NGINX, outsourced systems, container images, ingress, reverse proxies and embedded versions; check HTTP/2 and vendor patches.
- 02
Update and restart
Update each affected branch using vendor and distribution guidance. Restart and confirm old workers have exited. If updates must wait, evaluate temporarily disabling HTTP/2 and test compatibility.
- 03
Limit entry and monitor
Restrict unnecessary access; monitor unusual long connections, high HTTP/2 volume, worker crashes and unexpected child processes. These signals alone do not prove exploitation.
See what to doResources: 1
TWISA member solutions
Limit entry and monitor
Restrict unnecessary access; monitor unusual long connections, high HTTP/2 volume, worker crashes and unexpected child processes. These signals alone do not prove exploitation.
- 04
Investigate and retest
Preserve process, application, proxy and network logs; investigate anomalies, retest after updating and review backend privileges and data.
See what to doResources: 1
TWISA member solutions
Investigate and retest
Preserve process, application, proxy and network logs; investigate anomalies, retest after updating and review backend privileges and data.
Within an authorized scope, perform asset discovery, vulnerability scanning, penetration and API testing, attack simulation and retesting. Verify WAF requests, alerts, actual blocking and backend results. WAF and upstream proxy behavior depends on deployment and does not replace updates; separately assess internal and local risks.
Inventory public and internal NGINX, outsourced systems, container images, ingress, reverse proxies and embedded versions; check HTTP/2 and vendor patches.
Allow authorized traffic and apply controls at each layer.
- 01
External entry
Inventory and limit exposure
Inventory exposure and dependenciesInventory and limit exposure
- 02
NGINX and proxy
Update packages and embedded components
Update and restartUpdate packages and embedded components
- 03
Internal services
Segment and restrict privileges and data
Limit entry and monitorSegment and restrict privileges and data
Feed logs and traffic data from each layer into monitoring and response.
- Investigate and retestCorrelate network, NGINX, host and service records to investigate and recover.
Inventory exposure and dependencies
Inventory public and internal NGINX, outsourced systems, container images, ingress, reverse proxies and embedded versions; check HTTP/2 and vendor patches.
Limit entry and monitor
Restrict unnecessary access; monitor unusual long connections, high HTTP/2 volume, worker crashes and unexpected child processes. These signals alone do not prove exploitation.
Investigate and retest
Preserve process, application, proxy and network logs; investigate anomalies, retest after updating and review backend privileges and data.
Contact TWISA members for more information about their products and services.
Alert updated: 2026/11/19. Refer to the vendor for current product support and remediation status.