---
id: CVE-2025-0453
aliases:
  - GHSA-49m6-vrr9-2cqm
  - BIT-mlflow-2025-0453
  - PYSEC-2026-1637
title: MLflow Uncontrolled Resource Consumption vulnerability
summary: MLflow Uncontrolled Resource Consumption vulnerability
severity: medium
cvss: 5.9
cvssVector: 'CVSS:3.0/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H'
vendor: mlflow
product: mlflow
ecosystem: pip
affected:
  - mlflow <= 2.17.2
published: '2025-03-20'
updated: '2026-07-07'
source: OSV
sourceUrl: 'https://osv.dev/vulnerability/GHSA-49m6-vrr9-2cqm'
references:
  - url: 'https://nvd.nist.gov/vuln/detail/CVE-2025-0453'
  - url: 'https://github.com/mlflow/mlflow'
  - url: 'https://huntr.com/bounties/788327ec-714a-4d5c-83aa-8df04dd7612b'
tags:
  - osv
  - pip
epss: 0.11033
epssPercentile: 0.95756
ingestedAt: '2026-07-08T18:25:45.450Z'
---

## Overview

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

- `mlflow <= 2.17.2`

## Remediation

Refer to the advisory for the patched release.