2 views collected around this technical thread.
This article explains how Baidu's vertical search platform tackled massive data growth and scaling challenges by redesigning its data management system, introducing elastic scheduling, decoupling ETCD access, implementing auto‑scaling, and advancing shard expansion to improve performance, stability, and cost efficiency.
This document describes a comprehensive capacity‑planning solution that predicts traffic‑peak impacts for hotel services, automatically estimates required CPU resources, creates timed scaling tasks, and evaluates performance using detailed metrics, thereby improving operational efficiency and reducing manual effort during events such as exam‑ticket printing and holiday travel surges.
This article presents a hybrid‑cloud solution that uses ACK Edge and KServe to dynamically allocate on‑premise and cloud GPU resources for large‑language‑model inference, addressing tidal traffic patterns, reducing costs, and ensuring high availability through elastic scaling and custom scheduling policies.
The article explains how elastic architecture, through auto‑scaling and failover mechanisms, dynamically adjusts resources and ensures continuous service during traffic spikes and component failures, improving cost efficiency, reliability, and operational stability for modern cloud‑based applications.
The article explains how to use throttling together with auto‑scaling to control resource consumption of cloud applications, prevent tenant overload, handle traffic bursts, and ensure service‑level agreements while optimizing costs.
To overcome native HPA limits and business‑specific constraints in a fully containerized, cloud‑native Kubernetes environment, we implemented a dual‑threshold water‑level and scheduled scaling engine, hybrid‑cloud ClusterAutoScale, mixed‑deployment resource prioritization, and comprehensive Prometheus‑based observability, achieving higher utilization, lower costs, and a roadmap toward deeper optimization and AIOps.
The Taobao Open Platform adopts a cloud‑native, multi‑tenant architecture that abstracts infrastructure, isolates tenants via independent or shared switch‑plus‑security‑group schemes with dual ENI pod networking, and leverages Kubernetes auto‑scaling to simplify onboarding, cut operational costs, and enable future low‑code and FaaS extensions.
ByConity is a cloud‑native, storage‑compute separated data warehouse engine that supports multi‑tenant isolation, high performance, and elastic scaling; this guide explains its three‑layer architecture, hardware requirements, Helm‑based Kubernetes deployment, dynamic scaling, and practical SQL testing steps.
This article details Hupu's massive sports‑traffic environment, the design and implementation of the Flex auto‑scaling platform, its architecture, core functions such as resource statistics, node and pod scaling, scenario scheduling, and the performance optimizations that enable rapid, cost‑effective scaling across multi‑cloud Kubernetes clusters.
The article presents high_service, a high‑performance inference framework that boosts GPU utilization in multimedia AI services by separating CPU‑heavy preprocessing from GPU inference, employing priority‑based auto‑scaling, multi‑tenant sharing, and TensorRT‑accelerated models to eliminate GIL bottlenecks, reduce waste, and adapt to fluctuating traffic, with future work targeting automated bottleneck detection and further CPU‑GPU offloading.
The session detailed how Tencent’s container experts boost Kubernetes cluster utilization by correcting pod resource requests, employing two‑level auto‑scaling, dynamic over‑commit, adaptive scheduling and eviction, and using HPA/EHPA/VPA, achieving up to 38.7% node usage and roughly 60% cost savings in real‑world workloads.
Knative is a CNCF‑incubated, cloud‑native serverless framework on Kubernetes that combines Build, Eventing, and Serving components—featuring a Knative Pod Autoscaler that can scale pods to zero—offering improved resource utilization, rapid traffic response, and developer productivity despite modest performance overhead.
This article explains the concept of elastic auto scaling in cloud services, describes typical scenarios such as high‑elastic web apps and compute‑intensive workloads, details the four‑layer architecture and workflow, and outlines functional features, stability improvements, and future optimization directions.
Shilong Fei from Xiaomi Data Platform presents an in‑depth exploration of elastic scheduling for Hadoop YARN, covering background, design of resource pools, auto‑scaling architecture, challenges such as job stability and user transparency, achieved cost reductions, and future plans for further optimization.
Shopee reduces data‑center carbon emissions by over 40,000 tons annually through three 2021 green‑computing technologies—Overcommit resource oversubscription, mixed‑model Colocation of latency‑sensitive and batch workloads, and enhanced Auto Scaling that leverages global metrics to cut machine usage and improve resource efficiency.
This article presents the design and implementation of Flink task auto‑scaling, covering background, manual and automatic scaling mechanisms, architecture with RescaleCoordinator, persistence via Zookeeper and HDFS, scaling policies for parallelism, CPU and memory, and future plans for fine‑grained and time‑based resource adjustments.
The article describes Liulishuo's Cloud Infra team's end‑to‑end automation of cloud resource provisioning, scaling, and cost governance using Terraform, a custom Luban platform, GitLab CI/CD, and chat‑bot integrations, highlighting the architectural design, implementation steps, and measurable benefits for both operations and business teams.
TDSQL‑C Serverless separates compute and storage, delivers instant elastic scaling for MySQL and PostgreSQL, charges per‑second usage, pauses and stops billing when idle, and supports low‑frequency, archival, development, and micro‑service workloads with a ~2‑second cold‑start.
This article explains how to achieve high availability for stateless services by employing redundancy, vertical and horizontal scaling, various load‑balancing algorithms (random, round‑robin, weighted, least‑connections, source‑hash), and automatic scaling techniques in cloud‑native environments, while also covering performance monitoring and CDN/OSS usage.
This article details Dada Group’s implementation of an intelligent elastic scaling architecture that automatically adjusts capacity during peak promotions and low‑traffic periods, improving delivery reliability, reducing costs, and supporting multi‑cloud and multi‑runtime environments through sophisticated monitoring and auto‑scaler mechanisms.