Share GPUs across teams or isolate them?
Wrong answer = wasted hardware or frustrated engineers. We're breaking down the decision framework live on March 5.
Register now: vcluster.com/multitenancy...
#AI #GPU #Kubernetes
At KubeCon EU, JPMC, NVIDIA, and vCluster Labs take the stage to discuss running Kubernetes at enterprise and AI scale.
Spots are limited. Register now: luma.com/28cq1bmb
#KubeConEU #Kubernetes #PlatformEngineering
Flat pod networks break multi-tenancy. Shubham Katara shows how to combine vCluster + Cilium Multi-Pool IPAM for namespace-scoped PodCIDRs. Clean firewall rules, zero IP waste.
Read here: medium.com/@shubham.kat...
#PlatformEngineering #Cilium #Networking #DevOps
Don't miss this session at NVIDIA GTC 2026, Lukas Gentele joins experts from NVIDIA and Netris to break down production-ready, multi-tenant GPU infrastructure for AI workloads.
📅 March 19 | 3:00 p.m. PST
Session details: www.vcluster.com/events/nvidi...
#NVIDIAGTC
When you're ready for production, move your vcluster.yaml config into Helm values and deploy on a real Kubernetes cluster. Same vCluster. Different runtime. Full walkthrough by @artemlajko.bsky.social : medium.com/@artem_lajko...
Need more nodes? Attach an external VM over Tailscale VPN with one curl command. Node-to-node communication is handled automatically. The pod actually ran on the external node, LoadBalancer routed traffic through it correctly.
What you get out of the box: built-in LoadBalancer without MetalLB. Pause and resume clusters. Secure sharing, invite a teammate into your running local cluster via kubeconfig or the vCluster Platform. They connect. They debug. On your actual environment.
vCluster in Docker (vind) changes this. Three commands. Less than three minutes. No homelab. No cloud costs. A real Kubernetes control plane running inside Docker.
KinD clusters rarely survive longer than a demo. Every setup change means rebuilding from scratch. No real LoadBalancer. No collaboration. It's local by design, and that's the problem.
#Kubernetes #DevOps #vCluster
🚀 Platform v4.7 and vCluster v0.32 are live!
Focused on resilience and day-2 ops. Debug Shell for embedded etcd troubleshooting, Virtual Cluster Status Page, expanded Postgres/MySQL backing store, and K8s v1.35 support.
Full release notes: vcluster.com/changelog
Heading to NVIDIA GTC 2026!
vCluster is a Silver Sponsor. If you're attending, come say hi at Booth #206.
📅 March 16-19 | 📍 San Jose, CA
More details: www.vcluster.com/events/nvidi...
#NVIDIAGTC #GTC2026 #Kubernetes
AI infrastructure: all on-prem vs all cloud is a false binary.
Winning pattern: intelligent placement
Sensitive jobs → private GPU
Elastic workloads → cloud
Factors: data sensitivity, size, latency, compliance
👉️ Read here: www.vcluster.com/blog/pragmat...
#GPU #Kubernetes #MLOps
Why it's faster:
✅️ One Docker container per node instead of nested containerd.
✅️ Images proxied from local Docker daemon.
If those extra seconds are breaking your flow, give vind a try: github.com/loft-sh/vind
#Kubernetes #Docker #OpenSource #DevTools #CloudNative
We benchmarked vind vs kind. vind creates clusters 2-3x faster (8-12s vs 25-30s for single node), loads images instantly (vs 5-15s per image), and uses half the resources. Here's what we found 👇️
Join us this Thursday for a live demo: Automating GPU clusters with Auto Nodes + vCluster Platform Free Tier.
This is hands-on. Bring questions, watch the demo live, or share what you've already built.
Register here: www.linkedin.com/events/74256...
#GPU #Kubernetes #MLOps
vind gives you: instant Docker image sync, real LoadBalancer IPs, pause/resume clusters, in-place upgrades, and join remote GPUs via VPN. Open source. Works on Mac, Linux, Windows.
Try it: github.com/loft-sh/vind
#Kubernetes #Docker #OpenSource #DevTools #CloudNative
We've used kind for years. Solid tool. But small friction points kept adding up: manually importing images, no working LoadBalancers, destroying clusters to change versions. So we built vind (vCluster in Docker) to fix that. 👇️
Kubernetes nodes/proxy permission has an RCE exploit. No CVE, Kubernetes says "working as intended."
vCluster blocks it by filtering websockets at the proxy layer.
Full breakdown: website.vcluster.com/blog/vcluster-blocks-kubernetes-nodes-proxy-rce-exploit
#Kubernetes #CyberSecurity
We'll cover setup, tenancy models, resource limits, and real examples across platform engineering, CI, GPU workloads, and homelabs.
Live demos + Q&A.
Register: www.linkedin.com/events/74225...
#Kubernetes #CloudNative #PlatformEngineering #DevOps
Lots of questions about the Free tier since launch. How do I set up Auto Nodes? What can I build with 64 CPUs? How do tenancy models work?
We're doing a live walkthrough on Feb 5th to answer all of this. 👇️
Most GPU clusters can't scale because they're reactive, not systematic.
The 5-stage path from cluster to AI factory:
1. Basic Clusters
2. Advanced Scheduling
3. Multi-Tenancy
4. Platformization
5. AI Factory
Full breakdown: www.vcluster.com/blog/from-gp...
Most GPU clusters can't scale because they're reactive, not systematic.
The 5-stage path from cluster to AI factory:
1. Basic Clusters
2. Advanced Scheduling
3. Multi-Tenancy
4. Platformization
5. AI Factory
Full breakdown: www.vcluster.com/blog/from-gp...
One cluster per workload is a security myth that doesn't scale.
When you hit 100+ clusters, you aren't just scaling isolation; you’re scaling configuration drift and replicating vulnerabilities.
Read why: website.vcluster.com/blog/kuberne...
#PlatformEngineering #Kubernetes #CloudSecurity
Running GPU workloads in Kubernetes? Learn how to deploy virtual clusters on GPU-enabled GKE for true multi-tenant isolation. Complete guide covers GKE setup, resource allocation & PyTorch testing.
Read the full post here: website.vcluster.com/blog/deploy-...
#Kubernetes #MLOps #GPU
Help us shape vCluster's 2026 roadmap.
We're running our Annual User Survey and would love your feedback.
Takes 5-8 minutes, and you'll get a $22 vCluster Store gift card as a thank-you.
👉 forms.gle/VAqpTnm7GVse...
Your insights help us build a better vCluster for everyone.
Kubernetes for AI workloads is a mess. GPU scarcity, poor utilization, and fragmented deployments.
@saiyampathak.com is covering how to tackle this at the @oreilly.bsky.social Infrastructure & Ops Superstream on January 20th.
👉 www.oreilly.com/live-events/...
📦 871K downloads
⭐ +3.9K GitHub stars
🚀 43 releases
Virtual clusters are becoming a production standard for multi-tenancy, AI/ML, and CI/CD. Better isolation than namespaces. Lighter than separate clusters.
Thanks for building with us. 2026 loading... 🚀
#vCluster #Kubernetes
Celebrating our 2025 community - spotlight #5: Fabrizio Veritas
Distributed systems expert and #Kubernetes enthusiast bringing expertise in cloud-native architectures and platform engineering.
Thank you for being part of the community! 🥰
#vCluster #PlatformEngineering
📚 GPU-Enabled Platforms on Kubernetes - eBook:
vcluster.com/ebook/gpu-en...
🎓 Educational Series (4-part webinars):
vcluster.com/gpu-enabled-...
📝 Understanding Kubernetes Multi-Tenancy:
vcluster.com/blog/underst...
Interested in contributing as well? 👇️
github.com/loft-sh/vclu...
Celebrating our 2025 community - spotlight: @danielepolencic.com who co-authored our GPU-Enabled Platforms on K8s eBook, led educational series on GPU multi-tenancy, and helped thousands with his sessions and blogs.
Thank you for an incredible year!
See his work in the thread 👇
