Skip to content
Threat Feed
medium advisory

Next.js Cache Components Vulnerable to Denial-of-Service via Connection Exhaustion (CVE-2026-44579)

Next.js applications using Partial Prerendering through Cache Components are vulnerable to connection exhaustion (CVE-2026-44579), where crafted POST requests to a server action trigger a request-body handling deadlock, consuming server capacity and leading to denial of service.

A denial-of-service vulnerability (CVE-2026-44579) exists in Next.js applications that utilize Partial Prerendering through the Cache Components feature. This flaw allows an attacker to exhaust server connections by sending specially crafted POST requests to a server action. The malicious requests trigger a deadlock in request-body handling, causing connections to remain open for an extended period. This leads to the consumption of file descriptors and server capacity, ultimately denying service to legitimate users. The vulnerability affects Next.js versions 15.0.0 up to 15.5.16 and 16.0.0 up to 16.2.5. The fix implemented involves treating the Next-Resume header as internal-only and stripping it from untrusted incoming requests to prevent external manipulation.

Attack Chain

  1. The attacker identifies a Next.js application using Partial Prerendering and Cache Components.
  2. The attacker crafts a malicious POST request targeted at a server action endpoint.
  3. The malicious POST request includes a Next-Resume header, intended to exploit the partial prerendering mechanism.
  4. The Next.js application, upon receiving the crafted request, enters a deadlock state while processing the request body.
  5. The connection remains open and consumes server resources, including file descriptors.
  6. The attacker sends multiple such malicious POST requests concurrently, amplifying the resource consumption.
  7. The server’s capacity to handle new connections diminishes due to the exhausted resources.
  8. Legitimate users are denied service as the server is unable to accept new connections or process their requests, resulting in a denial-of-service condition.

Impact

Successful exploitation of CVE-2026-44579 leads to a denial-of-service condition, rendering Next.js applications unavailable to legitimate users. The number of victims is directly proportional to the attacker’s ability to send concurrent malicious requests and the server’s capacity to handle connections. Sectors reliant on Next.js applications for critical services, such as e-commerce, content delivery, and web applications, are particularly vulnerable. The vulnerability can severely impact business operations, causing financial losses, reputational damage, and disruption of services.

Recommendation

  • Upgrade Next.js to version 15.5.16 or later for the 15.x branch, or version 16.2.5 or later for the 16.x branch, to incorporate the fix for CVE-2026-44579.
  • As a temporary workaround, block incoming requests containing the Next-Resume header at the edge to prevent exploitation until the upgrade can be performed (see Overview).
  • Deploy the Sigma rule Detect Suspicious Next-Resume Header to identify potential exploitation attempts by monitoring for the presence of the Next-Resume header in incoming HTTP requests.
  • Monitor web server access logs for a high volume of POST requests to server action endpoints, which could indicate an attempted denial-of-service attack, and correlate with the Next-Resume header to refine detection.

Detection coverage 2

Detect Suspicious Next-Resume Header

high

Detects HTTP requests containing the Next-Resume header, which is now considered internal-only in Next.js versions >=15.5.16 and >=16.2.5 and should not be present in external requests, indicating a potential CVE-2026-44579 exploitation attempt.

sigma tactics: impact techniques: T1499 sources: webserver

Detect High Volume of POST Requests to Server Action Endpoints

medium

Detects a high volume of POST requests to common server action endpoints, which could indicate a denial-of-service attack targeting specific Next.js functionalities.

sigma tactics: impact techniques: T1499 sources: webserver

Detection queries are available on the platform. Get full rules →