The 5th Annual State of FinOps survey conducted from within the FinOps Foundation community of large cloud spenders is now available. Here you'll find the takeaways from the survey which surveyed large companies responsible for over $69bn of cloud spend... read more  

Windows support for Virtual Machines Node Pools now available in Azure Kubernetes Services. Azure Kubernetes Services directly manages provisioning and bootstrapping of each node. Virtual Machines node pools support adding multiple VM SKUs of the same family to a single node pool. Public preview fea.. read more  

Prometheusoperates as a solitary,pull-drivenmonitoring system, archiving data in a refined format with PromQL for queries.Memory and local diskplay the role of custodians for data, yet they facescalability hurdlesamidcardinality explosion. Linking multiple servers can somewhat amelioratevisibility, .. read more  

TheCNCF awarded CubeFS graduation, a standoutdistributed storage systemcompatible with POSIX, HDFS, and S3, boasting a staggering1,900% surge in userssince its inception into the community. Initially conceived in 2017 by JD.com, it shows vigorous momentum, and the CNCF presented an ambitious roadmap.. read more  

Gloo Gateway, now named'kgateway'and generously gifted to CNCF at KubeCon NA 2024, holds the crown as a premierEnvoy-based gatewaywith impressive deployment. The project wrestled withserialization bottlenecksbecause of bloated Proxy resources. With finesse, the team conquered dependency management c.. read more  

Helmwields a templating engine to reign over Kubernetes applications, letting users dive into version control and rollbacks. Meanwhile,Kustomizerevels in its overlay-based method for YAML file customization, embedding itself seamlessly withkubectl. When they join forces, Helm orchestrates releases, .. read more  

Implemented a distributed speed test solution across all store systems to monitor network usability and speed. Created a speedtest-api-service in Golang with client and server components to schedule and perform speed tests. Utilized Grafana dashboards to display metrics for speed, latency, packetlos.. read more  

This post explores the use of efficient GPU orchestration for distributed training in MLOps, highlighting how GPUs can significantly boost performance at scale. It delves into key technical considerations such as system setup, orchestration strategies, and performance optimization for scaling modern.. read more  

Dapr Agentssimplifies AI agent creation, utilizing LLMs to craftsecure multi-agent systemswithstateful workflowsandtool integrationstailored for enterprise environments. It harnessesvendor-neutralframeworks, scales adeptly, and melds effortlessly withover 50 data sources, fosteringresiliencyandcost-.. read more  

GKE's Compute Class feature in Google Kubernetes Engine (GKE) defines node profiles for specific workloads, allowing users to prioritize configurations for performance, cost, and efficiency. Custom compute classes provide granular control over infrastructure choices, improving cost efficiency and re.. read more