August Product Update

Aaron Smallberg -

🔍 What's New in Plural

Cluster, Node, Namespace, and Inventory Visibility

This month, we've expanded the infrastructure visibility you already get with Plural to allow you to easily access and understand inventory metrics for your clusters, including number of nodes, pods, namespaces, and availability zones.

These new capabilities integrate seamlessly with Plural's unified dashboard, offering teams more detailed infrastructure cataloging and predictive resource forecasting across their entire Kubernetes fleet.

Support for OpenShift Clusters

​​Plural has expanded its platform compatibility to include support for Red Hat OpenShift clusters, enabling organizations running OpenShift environments to use Plural's fleet management capabilities alongside their existing Kubernetes infrastructure.

This expanded compatibility allows teams to manage mixed environments with both standard Kubernetes and OpenShift clusters through Plural's unified control plane.

🏢 Company Updates

AWS Summit

We’re sponsoring the AWS Summit in LA next month. Catch us on September 17, 2025 at the Los Angeles Convention Center - South. We'll be by the AWS AI Hub, Booth #639

Kubernetes Meetup

We had an action-packed meetup this week, with demos from Runhouse and MetalBear

Runhouse demoed reinforcement learning & how to do it on Kubernetes:

💡 RL workloads are complex relative to the model pre-training that dominated AI research in the past several years. This is because it mixes training, inference, and evaluation into a single workload rather than being a single long-running, fixed-size execution.

💡 Kubernetes is the best place to do ML because it enables great bin packing, resource sharing, reproducibility, portability, and plugs into a good observability ecosystem.

💡 Runhouse's product, Kubetorch, enables programmatic control of separately autoscaling sub-services, each running on different images.

MetalBear demod how to tighten the Kubernetes feedback loop with remocal development: 

🔄 Testing code changes in K8s doesn’t need to be lengthy or painful. You can test code changes in K8s instantly from your laptop.

🔄 Significant cloud cost savings come from using mirrord together as a team to share a stable staging environment rather than spinning up ephemeral environments or “dev spaces” for daily development work.

🔄 It’s easy to get started with MetalBear's mirrord. You can check out the OSS at https://github.com/metalbear-co/mirrord

Our CTO Michael Guarino then followed up with a demo of Plural's new AI DevOps co-pilot that can provide:

🤖  Automatic insights and app code-level troubleshooting
🤖  Natural Language ops to query infrastructure
🤖  Agentic workflow for DevOps

Plural Blog

We just published The Enterprise Scaling Challenge: Taming Multi-cluster, Multi-team Kubernetes. In this post, we explore how to overcome challenges like downtimeless upgrades, secure access, and faster troubleshooting, even as you scale your Kubernetes infrastructure to multiple clusters.