CVE-2026-133200

OpenShift AI TrustyAI streaming requests could exhaust workers

Red Hat · Red Hat OpenShift AI · rhoai/odh-trustyai-service-rhel9

First documented disclosure · Review checkpoint

Summary

Disconnected clients could leave blocking streaming responses consuming worker threads.

Severity and exploitation evidence

Medium · 5.9/10 · CVSS 3.1. Severity source.

Exploitation not established: Reviewed public sources describe the weakness, a correction or a controlled demonstration; malicious exploitation against users is not established.

Exploitation evidence source

Technical severity is not a measure of observed attacks or of risk to your deployment. Attempts, demonstrations, and successful compromise retain the qualifications in the source.

How the weakness works

quarkus-vertx-utils does not stop processing the streaming output after client disconnection.

Conditions required

The deployed service exposes blocking, long-lived or infinite REST streams; ordinary non-streaming endpoints are outside the described trigger.

Potential impact

Worker-pool exhaustion and denial of service.

Affected versions and configurations

Red Hat identifies the TrustyAI service package as affected; a complete package-version range is not supplied.

Fix and mitigation status

No fixed TrustyAI package identified; Red Hat status is Fix deferred.

Practical next steps

Track the package-specific Red Hat correction and limit exposure or resource use of affected blocking streams. Red Hat publishes no generally applicable mitigation.

Operating lesson

Check cancellation and resource limits at streaming boundaries.

Evidence and limits

Included for an explicitly affected OpenShift AI component, not every Quarkus consumer. Red Hat’s public timestamp is November 19 15: 01: 24Z; its vendor severity is Moderate and draft CVSS 3.1 is 5.9. CNA publication is 2026-11-19; this is distinct from the earliest reviewed public source and correction dates. redhat assessments: CVSS 3.1 5.9.

Sources