July 3rd, 2024

The Trouble with Tar: Introducing High Availability, Zone-Aware Load Balancing

Topology Aware Routing (TAR) in Kubernetes poses challenges by restricting traffic within zones, affecting system reliability. Linkerd's HAZL offers dynamic load balancing across zones for high availability and cost reduction, addressing TAR limitations.

Read original articleLink Icon
The Trouble with Tar: Introducing High Availability, Zone-Aware Load Balancing

In a blog post by William Morgan, the challenges of Topology Aware Routing (TAR) in Kubernetes are discussed. TAR restricts traffic within availability zones to save costs but can reduce system reliability by preventing pods in one zone from compensating for failures in others. While TAR is resilient to certain failures detected by liveness and readiness probes, it cannot handle in-band failures like application errors or broken dependencies. To address these limitations, service meshes like Linkerd offer in-band health checking, fine-grained traffic control, and the ability to handle cross-cluster traffic. Buoyant Enterprise for Linkerd introduces High Availability Zonal Load Balancing (HAZL) as a solution. HAZL dynamically balances traffic within and across zones to maintain high availability while reducing costs in multi-zone Kubernetes clusters. By leveraging features like in-band health checking and fine-grained traffic control, HAZL aims to provide a more flexible and reliable alternative to TAR for Kubernetes adopters seeking optimal cloud cost management and system reliability.

Related

Multiple Regions, Single Pane of Glass

Multiple Regions, Single Pane of Glass

WarpStream implements a hub-and-spoke model to provide highly available resources across regions. They use a push-based replication strategy with "contexts" for metadata distribution, prioritizing availability over consistency.

My First Kubernetes: k3s 'cluster' on 3 Orange Pi Zero 3's

My First Kubernetes: k3s 'cluster' on 3 Orange Pi Zero 3's

Setting up a Kubernetes cluster with Orange Pi Zero 3 boards involves control and worker nodes, deploying apps, and overcoming ARM64 image compatibility challenges. Detailed steps, including Dashboard setup and user management, are shared, offering practical insights for Kubernetes enthusiasts.

Bad habits that stop engineering teams from high-performance

Bad habits that stop engineering teams from high-performance

Engineering teams face hindering bad habits affecting performance. Importance of observability in software development stressed, including Elastic's OpenTelemetry role. CI/CD practices, cloud-native tech updates, data management solutions, mobile testing advancements, API tools, DevSecOps, and team culture discussed.

Cubernetes

Cubernetes

Justin Garrison built "Cubernetes," a visually appealing Kubernetes hardware lab for training and content creation. The $6310 setup included unique parts like Mac Cube cases and LP-179 computers with Intel AMT support. Creative solutions like 3D printing and magnetic connectors were used. Lights were controlled by attiny85 and Raspberry Pi Pico for visualizations. The project prioritized functionality and education.

For the Love of God, Stop Using CPU Limits on Kubernetes

For the Love of God, Stop Using CPU Limits on Kubernetes

Using CPU limits on Kubernetes can lead to CPU throttling, causing more harm than good. Setting accurate CPU requests is crucial to avoid throttling. Memory management best practices are also discussed, along with a tool for resource recommendations.

Link Icon 2 comments