Uncontrolled Resource Consumption
CVE-2025-48956
Summary
The vLLM is an inference and serving engine for large language models (LLMs). In versions 0.1.0 through 0.10.1, a Denial of Service (DoS) vulnerability can be triggered by sending a single HTTP GET request with an extremely large header to an HTTP endpoint. This results in server memory exhaustion, potentially leading to a crash or unresponsiveness. The attack does not require authentication, making it exploitable by any remote user.
- LOW
- NETWORK
- NONE
- UNCHANGED
- NONE
- NONE
- NONE
- HIGH
CWE-400 - Uncontrolled resource consumption
An uncontrolled resource allocation attack (also known as resource exhaustion attack) triggers unauthorized overconsumption of the limited resources in an application, such as memory, file system storage, database connection pool entries, and CPU. This may lead to denial of service for valid users and degradation of the application's functionality as well as that of the host operating system.
References
Advisory Timeline
- Published
