In this post, we explore KServe, a model inference platform on Kubernetes designed for scalability. Building on our previous Kubeflow guide, we detail how to set up your first KServe endpoint, make predictions, and troubleshoot common issues. Follow our step-by-step instructions to seamlessly integrate KServe with your Kubeflow environment and enhance your machine learning deployment process.

This comprehensive guide walks you through integrating Kong, a powerful unified API platform, into your home lab environment. Starting with Docker installation and a custom Kong image, it covers PostgreSQL deployment, MetalLB setup for load balancing, and Kong service and route configuration. The post concludes with troubleshooting tips and instructions for deploying your customized Kong image in a K3s cluster. This step-by-step tutorial empowers you to efficiently manage APIs in your home lab using Kong.

I've successfully set up Helm as the package manager for my Raspberry Pi Kubernetes Cluster, enabling easy discovery, sharing, and use of Kubernetes software. After installing Helm, I set up the NGINX Ingress Controller using Helm for external access. Additionally, I configured MetalLB as a layer 2 load balancer to manage external IP addresses. Revisiting my Gitea setup, I updated the service configuration to leverage MetalLB, enhancing the overall scalability and efficiency of my Kubernetes environment on Raspberry Pi. Helm proves to be a valuable tool for managing Kubernetes packages effortlessly.