Vulnerability · Published Mar 20, 2025 · Assigned by @huntr_ai

CVE-2025-0453Denial of Service through Batched Queries in GraphQL in mlflow/mlflow

Medium 5.9

In plain English

A medium-severity security flaw in mlflow/mlflow from mlflow. Proof-of-concept exploit code is public. The chance it is exploited in the next 30 days is estimated at 11.3% (higher than 96% of all known vulnerabilities).

Severity (CVSS 3.0)5.9Medium
Exploit odds (30 days)11.3%top 4.0% of all CVEs
Exploited in the wildNo reportsCISA SSVC: poc
Attack toolingNoneno public tooling known

Official description

In mlflow/mlflow version 2.17.2, the `/graphql` endpoint is vulnerable to a denial of service attack. An attacker can create large batches of queries that repeatedly request all runs from a given experiment. This can tie up all the workers allocated by MLFlow, rendering the application unable to respond to other requests. This vulnerability is due to uncontrolled resource consumption.

Affected packages & fixed versions

EcosystemPackageFixed inAdvisory
PyPImlflowno fix listedGHSA-49m6-vrr9-2cqm
PyPImlflowno fix listedPYSEC-2026-1637

Affected products (as reported)

VendorProductVersions
mlflow mlflow/mlflow unspecified – through latest

Weakness: CWE-410 · Vector: CVSS:3.0/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H

References