CVE-2026-105759: vLLM: Unbounded Prometheus label cardinality from attacker-controlled HTTP method tokens in the vLLM Rust frontend metrics middleware (unauthenticated denial of service)
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, the Rust frontend's trackhttpmetrics middleware records the raw HTTP method token as a Prometheus label for requests reaching registered routes. An unauthenticated attacker can send unique arbitrary method tokens to unguarded routes such as /tokenize, causing Prometheus's Family::getorcreate function to permanently create counter and histogram label sets. Those label sets increase process memory usage and enlarge the /metrics response until the service or monitoring path is exhausted. This issue is fixed in version 0.30.0.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade
vLLMto a version that resolves this vulnerability.Fixed in 0.30.0
Event History
Frequently Asked Questions
Which deployments are exposed?
vLLM versions before 0.30.0 are affected when using the Rust frontend and requests can reach registered routes. Unguarded endpoints such as /tokenize are specifically identified as reachable attack targets.
What does an attacker need to exploit this issue?
An attacker does not need authentication or user interaction. They need network access to send requests using many unique arbitrary HTTP method tokens to a reachable registered route.
What signs indicate active exploitation or impact?
The Prometheus metrics will accumulate permanently created counter and histogram label sets for distinct HTTP method tokens. This can appear as growing process memory consumption and an increasingly large /metrics response, eventually exhausting the service or monitoring path.
What version contains the fix?
The issue is fixed in vLLM version 0.30.0.