A k8s virtual kubelet that runs GPU jobs on RunPod.
-
Updated
Sep 6, 2025 - Go
A k8s virtual kubelet that runs GPU jobs on RunPod.
NATS-to-Kubernetes cloud-bursting controller for AI workloads. Publish JobDescriptors on a NATS bus, a Go controller applies them to a remote K8s cluster (e.g. NRP Nautilus) with CSMA/CA-style politeness backoff.
Addressing transient workload spikes with Cloud bursting in Apache Flink
Learn the AWS Plugin for Slurm v2 by deploying disposable, version-matched mock on-prem HPC clusters on AWS. Five generations (Rocky 8/9/10 + Ubuntu 22.04/24.04, Slurm 22.05/23.11/24.05). Terraform + Packer.
ASBA - Intelligent decisions for academic researchers using SLURM clusters to burst computational workloads to AWS EC2. Analyzes job history, optimizes resources, and recommends optimal instance types for cost-effective research computing.
This project automates a hybrid cloud setup with on-prem servers and AWS for scalability, high availability, and cost optimization. It features: ✅ HAProxy + Keepalived for load balancing & failover ✅ Cloud Bursting via AWS Auto Scaling ✅ Secure WireGuard VPN for hybrid connectivity ✅ Terraform + Ansible for full automation
Hard money-budget enforcement for Slurm jobs — reserves projected cost at submission, rejects unfundable jobs, settles actual runtime, refunds the rest. Hierarchical accounts, banked burst, multi-source funding.
Global Namespace for Hybrid HPC Clouds - Seamless data access across on-premises and cloud compute resources
Add a description, image, and links to the cloud-bursting topic page so that developers can more easily learn about it.
To associate your repository with the cloud-bursting topic, visit your repo's landing page and select "manage topics."