Skip to content

ray

Model Serving & MLOps

A unified framework for scaling AI and Python applications from a laptop to a cluster

Python Latest ray-2.55.1 · 1mo ago Security brief →

Features

  • Unified runtime that scales Python code across laptops, clusters, clouds, and Kubernetes
  • AI libraries for scalable ML datasets (Data), distributed training (Train), hyperparameter tuning (Tune), reinforcement learning (RLlib), and programmable serving (Serve)
  • Core abstractions: Tasks (stateless functions), Actors (stateful workers), and Objects (immutable shared values)

Security Response History

1 CVE
CVE Severity Disclosed Patched (this tool) vs Ecosystem Median
CVE-2023-4863 KEV high
CVSS 8.8
2023-09-13 2026-02-18 2y 5mo / median 2y 4mo

Recent releases

View all 4 releases →
ray-2.55.1 Bug fix

Minor fixes and improvements.

Full changelog
  • Fixes SSH connectivity issue in the ray-llm image (#62625 / #62718).
  • Upgrade apt packages in slim base (#62666 / #62717).
ray-2.55.0 Breaking risk
Breaking changes
  • local_mode support removed
  • Legacy BlockList, locality_with_output, and callback API removed
  • PyArrow 9.0 compatibility checks removed
Security fixes
  • jackson-databind upgraded from 2.16.1 to 2.18.6 (GHSA-72hv-8253-57qq)
Notable features
  • DataSourceV2 API with scanner/reader framework and file partitioning
  • Kafka datasink with confluent-kafka migration and datetime offset support
  • 2-phase commit checkpointing with trie recovery
ray-2.54.1 Bug fix

Hanging issue detector disabled to resolve pipeline performance degradation from blocking API calls.

ray-2.54.0 New feature
Notable features
  • New cluster autoscaler enabled by default
  • Ray Data checkpointing support
  • Queue-based autoscaling for Ray Serve TaskConsumer

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

About

Stars
42,682
Forks
7,604
Languages
Python C++ Java

Install & Platforms

Install via
pip

Community & Support

Beta — feedback welcome: [email protected]