PatchSiren cyber security CVE debrief
CVE-2026-105759 vllm-project CVE debrief
An inference and serving engine for large language models, vLLM, is vulnerable to a denial-of-service (DoS) attack. Prior to version 0.30.0, the Rust frontend's track_http_metrics 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::get_or_create 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.
- Vendor
- vllm-project
- Product
- vllm
- CVSS
- MEDIUM 5.9
- CISA KEV
- Not listed in stored evidence
- Original CVE published
- 2026-10-05
- Original CVE updated
- 2026-10-23
- Advisory published
- 2026-10-05
- Advisory updated
- 2026-10-23
Who should care
Defenders who are responsible for the security of systems using vLLM should care about this vulnerability. They should assess their exposure to this vulnerability and prioritize updating to version 0.30.0 or later to prevent this DoS vulnerability.
Why it matters
This vulnerability allows an unauthenticated attacker to cause a DoS attack on systems using vLLM prior to version 0.30.0. Defenders should prioritize updating to version 0.30.0 or later to prevent this vulnerability. They should also monitor their systems for unusual traffic patterns and assess their exposure to this vulnerability.
- Increased process memory usage
- Enlarged /metrics response size
- Potential service or monitoring path exhaustion
- DoS vulnerability allowing unauthenticated attackers to send unique arbitrary method tokens
Technical summary
The vulnerability is caused by the Rust frontend's track_http_metrics middleware recording the raw HTTP method token as a Prometheus label for requests reaching registered routes. This allows an unauthenticated attacker to send unique arbitrary method tokens to unguarded routes, causing Prometheus's Family::get_or_create function to permanently create counter and histogram label sets. These label sets increase process memory usage and enlarge the /metrics response until the service or monitoring path is exhausted.
Defensive priority
Defenders should prioritize updating to version 0.30.0 or later to prevent this DoS vulnerability. They should also monitor their systems for unusual traffic patterns and assess their exposure to this vulnerability.
Recommended defensive actions
- Update to version 0.30.0 or later
- Monitor systems for unusual traffic patterns
- Assess exposure to this vulnerability
- Review compensating controls for exposed systems while remediation is scheduled and verified
- Check relevant monitoring, detection, and logs for exposed assets that need extra review
- Track exceptions, retest remediated assets, and close the item only after evidence is documented
- Confirm whether affected product deployments exist in managed environments and assign an owner for follow-up
Evidence notes
The evidence for this vulnerability comes from the official CVE Program record, the NVD vulnerability detail page, and the source item from osv_dev. The CVE Program record provides the official CVE metadata, while the NVD vulnerability detail page provides a source-specific vulnerability assessment. The source item from osv_dev provides additional details about the vulnerability, including the affected versions and the fixed version.
Sources and references
Verified primary and authoritative sources
-
CVE-2026-105759 CVE Program record
Publisher, destination, and source semantics verified
URL: https://www.cve.org/CVERecord?id=CVE-2026-105759
CVE Program - Official CVE Program record with source-provided CVE metadata.
-
CVE-2026-105759 NVD vulnerability detail
Publisher, destination, and source semantics verified
URL: https://nvd.nist.gov/vuln/detail/CVE-2026-105759
NIST National Vulnerability Database - Official NIST NVD detail page and source-specific vulnerability assessment.
Supplemental references
-
PYSEC-2026-4325
Unverified legacy reference
URL: https://storage.googleapis.com/osv-vulnerabilities/PyPI/PYSEC-2026-4325.json
osv_dev
-
Source reference
Unverified legacy reference
URL: https://github.com/vllm-project/vllm/releases/tag/v0.30.0
Supplemental source
-
Source reference
Unverified legacy reference
URL: https://github.com/vllm-project/vllm/commit/3735c2d5f5248259482b9045c34fb7a8a3892352
Supplemental source
-
Source reference
Unverified legacy reference
URL: https://github.com/vllm-project/vllm/pull/56058
Supplemental source
-
Source reference
Unverified legacy reference
URL: https://github.com/vllm-project/vllm/security/advisories/GHSA-5fj9-pfhr-6j48
Supplemental source
Methodology and review provenance
AI-assisted synthesis based on stored public vulnerability evidence. System validation, approval state, and publication status do not by themselves establish human review of this revision. PatchSiren helps prioritize defensive review and does not prove exposure or remediation on any system.