Ray on GKE, cheaper storage, binding blocking and Valkey

The News

GKE

• Ray Add-on for GKE: GKE added a manager Add-on to deploy Ray on GKE. Ray is a popular AI python framework for writing and running AI applications. With this Add-on you can just tick a box and we will deploy and manage Ray for you.
• Optimize storage perf and costs with Hyperdisk: Hyperdisk is a Next-Gen network attached storage feature in Google Cloud. With Hyperdisk Storage Pools you can pre-purchase a collection of capacity, throughput, and IOPS that you can then provision to your applications as needed. When you place Hyperdisk block storage disks in storage pools, you can share capacity and performance among all the disks in a pool, optimizing for both operations and cost. Available from GKE 1.29.2.
• Block system bindings on GKE: Customers often misconfigure their clusters by creating overly permissive bindings to the system:authenticated. Which was called out by ORCA security in their blog post here and here. This can inadvertently grant broad access privileges, creating a potential security vulnerability. With this feature you can block system binding on GKE during cluster creation.
• GKE Entreprise Network Topology: Network Topology is a visualization tool that shows the topology of your network infrastructure for your VPC and your GKE enterprise clusters.

Google Cloud

• Cloud Run get deterministic URL’s: No more guessing. You can pre-determine your service URL before creating it. Keep in mind that this works only if the DNS length of 63 or less. the length is defined by the service-name length.

The Editorial

GKE

• Ten GKE networking problems and how to solve them: This guide explores ten common GKE networking problems and provides actionable steps to resolve them.
• Intro to Ray on GKE: An intro tutorial to get started with Ray on GKE.
• Disabling Public GKE Cluster with Custom Organization Policy Constraints: If you want to prevent people from creating GKE clusters with Public IP’s (You know for security reasons) you can use a customer organization policy. This tutorial shows you how.
• Connect from Google Kubernetes Engine (GKE) to AlloyDB for PostgreSQL: This tutorial describes how to set up a connection from an application running in Google Kubernetes Engine autopilot cluster to an AlloyDB instance.
• Pick the right AI infrastructure options in GKE: Recommendations that can help you maximize your serving throughput on NVIDIA GPUs on GKE.
• Choosing between self-hosted GKE and managed Vertex AI to host AI models: In this blog post, we equip developers, operations specialists, or IT decision-makers to answer the critical questions of "why" and "how" to deploy modern apps for LLM inference. We'll address the balance between ease of use and customization, helping you optimize your LLM deployment strategy.

Google Cloud

• Select a managed container platform: This article helps you select a proper manager container platform on Google Cloud that meets your application requirements.
• Memorystore for Valkey is now in preview: With the uncertainties around OSS licenses. Valkey was created as a fork for Redis. We announced support for Valkey in Memorystore in Preview.