Naftis is an open‑source web dashboard for Istio that lets users define custom task templates, manage services, view topology and metrics, and provides step‑by‑step Kubernetes deployment instructions, Docker images, and a complete developer guide for building and running the platform.
This article explains how to solve the problem of losing Jenkins job artifacts when using dynamic Kubernetes agents by mounting the slave workspace to the master via NFS, modifying pipeline code, and accessing the workspace through the Jenkins UI, complete with step‑by‑step instructions and troubleshooting tips.
Although Docker’s libnetwork/CNM offers a network plugin model, Kubernetes has chosen not to adopt it due to fundamental design mismatches, reliance on low‑level key‑value stores, security and scalability concerns, and a strategic shift toward the simpler, more portable CNI plugin framework.
This article provides a comprehensive overview of Kubernetes, explaining its purpose as an open‑source platform for automating container deployment, scaling, and management, detailing its advantages over traditional VM approaches, and outlining the core features, capabilities, and limitations of the system.
Alibaba’s PouchContainer, an internally built container runtime, combines LXC and Docker technologies to enable full‑scale, non‑intrusive containerization of online services, offering rich features such as systemd‑based rich containers, enhanced isolation, P2P image distribution, in‑place upgrades, and native Kubernetes integration.
The article provides an in‑depth technical overview of OpenStack Magnum, covering its community contributions, architecture, Nova and Heat Docker drivers, core concepts such as Bay and Baymodel, supported back‑ends like Kubernetes and Swarm, and the workflow for provisioning container clusters.
Manbang Group transformed thousands of mixed‑deployed Java microservice instances into Docker containers managed by Kubernetes, addressing resource isolation, manual fault handling, slow rollbacks, and DEV/QA release friction while implementing load balancing, service discovery, high availability, rolling upgrades, auto‑scaling, rapid deployment, and resource limits across a multi‑cluster cloud‑native architecture.
OpenStack can be integrated with Kubernetes either by deploying K8s on top of OpenStack—leveraging Magnum for rapid VM‑based clusters—or by tightly coupling K8s with OpenStack services such as Keystone, Neutron, Cinder, and Manila, each offering distinct benefits and trade‑offs.
This article presents a DNS‑based universal addressing solution that enables traditional SOA applications using interfaces such as Dubbo, HSF, and SOFA to run on a Service Mesh without code changes, by mapping interface names to Kubernetes ClusterIP through CoreDNS and leveraging Istio’s routing mechanisms.
This article provides a comprehensive technical guide on using Kubernetes and OpenShift for container‑cloud deployment, covering architecture, multi‑tenant isolation, security, load balancing, logging, monitoring, CI/CD, storage options, and best practices especially for financial‑industry microservice workloads.
This article explains how Kubernetes evolved from Google’s Borg research to become the leading cloud‑native container orchestration platform, detailing its master‑node architecture, core components such as kubelet and CRI, the role of Pods, Services, Deployments, and the declarative API model that enables scalable, flexible workload management.
The article reviews the evolution of Service Mesh and Dubbo, compares sidecar and SDK deployment models, discusses key technical challenges such as multi‑language support, performance, and operations, and explains how Dubbo integrates with Kubernetes to enable cloud‑native micro‑service architectures.
This article examines Youzan's journey to containerize its development and testing environments using Kubernetes and Docker, detailing the motivations, architectural decisions, network and isolation challenges, image integration, logging, load balancing, debugging, and the ongoing rollout to standard production environments.
Since its launch in 2016, JD's J-one unified release platform has evolved through multiple large‑scale promotions, introducing container‑based compilation, auto‑scaling, and a new image‑deployment capability that dramatically speeds up releases while improving success rates and providing a suite of automation tools for developers.
The 2018 Cloud Native Best Practices Summit in Beijing gathered over 200 industry leaders and experts who shared insights on cloud native trends, standards, Kubernetes, service mesh, product launches, and the formation of the Cloud Native Practice Alliance, marking a milestone for China's cloud native ecosystem.
Amid the rise of serverless Kubernetes, growing pod volumes, and real-time log demands, this article examines emerging log-handling challenges, evaluates traditional collection methods, and presents a comprehensive “Logtail + Log Service + Ecosystem” architecture that delivers high-throughput, reliable, and scalable logging for cloud-native environments.
This article examines OpenStack’s evolution over the past eight years, analyzes its current positioning against VMware and AWS, discusses the challenges of its “big tent” model and core services, and speculates on possible future paths for the platform from a user’s viewpoint.
Alibaba achieved full internal containerization by evolving from monolithic apps to the T4 container, integrating Docker images, and open‑sourcing PouchContainer, detailing the architectural shifts, resource isolation requirements, large‑scale deployment strategies, and lessons for building cloud‑native platforms.
This article explains how Kubelet obtains pod configurations, processes updates through its syncLoop, coordinates pod workers, and ultimately invokes the container runtime's SyncPod method to keep node pods aligned with the desired state.
This article compares the three major container orchestration platforms—Kubernetes, Docker Swarm, and Amazon ECS—examining their architecture, integration with cloud providers, deployment complexity, and suitability for different workloads to help readers choose the right tool for their needs.
This article explains why UCloud introduced a Service Mesh, how Istio was chosen and adapted for the UAEK Kubernetes platform—including IPv6 challenges, performance testing, and future improvements—providing practical insights for building reliable micro‑service infrastructures.
This guide explains how Kubernetes namespace-level resource management lets administrators set default CPU/memory requests, define minimum and maximum constraints, and enforce resource quotas, with step‑by‑step commands and YAML examples to create namespaces, ResourceQuota objects, and pods while handling quota limits.
The article details Alibaba’s Sigma scheduling and cluster management platform—its three‑layer architecture, data and state consistency strategies, real‑world case studies, Go‑based redesign, integration with Kubernetes APIs, and lessons on concurrency, high availability, and pod dispersion for massive Double 11 traffic.
The article describes how 58's search technology team standardized and modularized their vertical search services into a cloud-native platform, detailing the overall architecture, Kubernetes‑based deployment, service and deployment designs, stateful indexing challenges, and solutions for high‑availability and scalability.
Ant Financial tackles the challenge of managing dozens of Kubernetes clusters and over a hundred thousand worker nodes by employing a meta‑cluster with Kube‑on‑Kube and Node Operators, enabling automated lifecycle management, scaling, upgrades, and fault recovery for both master components and worker nodes.
The article explains why operating massive Kubernetes clusters is as challenging as building large systems, outlines three critical stability questions, shares real‑world data collection, visualization, and tooling practices, and provides concrete recommendations for high‑availability, monitoring, and performance optimization.
This article provides a comprehensive overview of Kubernetes, covering its definition, primary functions, core concepts such as Pods, Services, Replication Controllers and Labels, and detailed explanations of master components, Kubelet, and Proxy within the cloud‑native container orchestration ecosystem.
The article shares practical experience on operating massive Kubernetes clusters, focusing on three stability questions, data collection and visualization, and a suite of operational tools to achieve reliable, high‑availability services in production environments.
Istio 1.0, released on August 1, brings the first production‑ready service‑mesh platform with major enhancements in networking, policy & telemetry, out‑of‑process adapters, and security features such as mTLS and JWT, while introducing the IstioGateway to replace the legacy Ingress model.
This tutorial walks you through the complete process of manually installing a Kubernetes 1.9.9 master‑worker cluster on CentOS 7, covering environment preparation, binary downloads, configuration of etcd and Kubernetes components, creation of systemd services, and verification of the deployment.
Istio 1.0, now production‑ready, launches after a year of rapid growth, offering beta‑grade features such as multi‑cluster support, fine‑grained network APIs, incremental mutual TLS rollout, and enhanced Envoy performance, while an expanding ecosystem of observability, policy, and serverless projects backs its adoption.
This article outlines essential container operation best practices—including native logging, JSON log formatting, sidecar aggregators, stateless and immutable design, avoiding privileged containers, effective monitoring, health checks, non‑root execution, and careful image tagging—to help developers build secure, maintainable, and observable workloads on Kubernetes.
This article shares JD.com’s extensive experience redesigning Kubernetes for massive production use, covering custom DNS and load‑balancing, scaling clusters to ten‑thousand nodes, adapting controllers, building the Archimedes scheduler, and practical insights on resource isolation, deployment, and high‑traffic elasticity.
In a detailed Q&A session at ArchSummit 2018, Ant Financial and Alibaba UC experts discuss the fundamentals, challenges, implementation strategies, and future directions of Service Mesh and Kubernetes, covering performance, multi‑language support, open‑source efforts, and practical advice for developers and operators.
This article shares practical experience and best‑practice guidelines for operating large‑scale Kubernetes clusters, covering stability checks, component failure impact, recovery strategies, alerting mechanisms, data collection, visualization, and the suite of operational tools that help ensure reliable, high‑performance cloud‑native infrastructure.
The article describes JD’s experience of redesigning Kubernetes at massive scale, detailing the JDOS2.0 platform, custom DNS and load‑balancing, the Archimedes scheduler, API and controller optimizations, and operational lessons learned from running tens of thousands of nodes in production.
The article outlines ten major challenges faced when adopting microservice architecture—such as data synchronization, security, version control, discovery, data expiration, debugging, testing, monitoring, DevOps support, and fault tolerance—and presents practical solutions using cloud‑native tools and patterns.
TencentHub is a unified storage and DevOps platform that combines Docker images, Helm charts, and binary artifacts, using a Kubernetes‑based architecture with a Registry layer, COS‑backed storage engine, JWT token authentication, built‑in vulnerability scanning, and a YAML‑driven, containerized workflow engine for automated, scalable software delivery.
The talk details XtraPi’s journey from early PBS‑based supercomputers to a modern Kubernetes‑driven, multi‑cloud platform that uses Tencent Cloud TKE to run massive containerized drug‑discovery simulations, describing scaling strategies, image optimization, CI pipelines, checkpoint‑restart, and future serverless and bare‑metal enhancements.
This article explains the fundamental differences between Linux containers and virtual machines, covering resource usage, use‑case scenarios, portability, security considerations, and how to choose the right technology for various application and infrastructure needs.
Tencent Cloud’s Container Instance Service (CIS) offers a fully managed, serverless‑style Kubernetes platform that launches Docker images instantly with per‑second billing, VPC isolation, and elastic networking, while Clear Container (Kata Containers) provides lightweight VM‑based isolation for containers, enhancing security and flexibility.
This article explains how Kubernetes NetworkPolicy works with Calico, detailing the policy creation process, configuration templates, component versions, deployment steps, Calico‑felix packet handling, and a practical “deny all traffic” test case, concluding with operational considerations and reference links.
Tencent Cloud’s fully managed Kubernetes service, launched in 2016, delivers one‑click, VPC‑isolated cluster deployment with automated lifecycle, integrated monitoring, logging, storage (CBS/CFS) and CI/CD, custom components for metrics and storage, flat VPC networking, CSI drivers, and flexible master deployment models that simplify scaling, upgrades, and developer focus on applications.
The article recaps a Cloud+ Tech Salon session where Tencent Cloud’s one‑click Kubernetes service, the serverless Container Instance Service, and the Tencent Hub DevOps platform are detailed, illustrating how their integrated networking, storage, monitoring, and workflow features simplify large‑scale AI/ML and high‑performance scientific computing deployments in the cloud.
This article describes JD.com's JDOS Operations Platform, which enables two operators to manage millions of Docker and Kubernetes containers across massive clusters, detailing its architecture, regression analysis of scale, gossip‑based inspection system, intelligent alert convergence, and performance improvements for ultra‑large‑scale environments.
The article describes how JD.com built and operates the JDOS Operations Platform to manage a multi‑million‑container Docker and Kubernetes fleet, detailing the challenges of massive scale, the architectural components such as the configuration center, operation center, inspection system, gossip‑based communication, and an intelligent alerting system that together enable efficient, automated, and reliable large‑scale container operations.
This article walks through the design and implementation of a production‑ready continuous integration and delivery pipeline using Jenkins, Artifactory, Kubernetes, and Helm, covering pipeline architecture, artifact promotion, binary metadata linking, resource limits, multi‑container pods, Helm chart management, monitoring with EFK, and practical lessons learned.
This article provides a concise, question‑driven overview of Kubernetes, covering the roles of master and worker nodes, pod scheduling, data storage with etcd, service exposure, dynamic scaling, and how the various control‑plane components collaborate to manage a distributed container cluster.
This article explains the motivation behind the Kubernetes Container Runtime Interface (CRI), outlines its design, and provides a detailed walkthrough of PouchContainer’s CRI manager architecture, including pod creation, network configuration, and I/O stream handling, with code examples and configuration snippets.
The DevOps International Summit 2024 in Beijing, the sole China‑based global DevOps conference, brings together over 80 leading experts to showcase end‑to‑end practices—from Lean‑Agile, Continuous Delivery, SRE, and microservices to DevSecOps, AI‑driven tooling, and the new Research and Operations Integration Capability Maturity Model—through industry‑focused tracks, hands‑on training, and real‑world case studies across finance, telecom, retail and more.
This guide walks you through installing a Kubernetes (K8S) cluster on CentOS 7.3 in China, covering environment preparation, downloading required packages and images via local mirrors, configuring Docker and the network, initializing the master node with kubeadm, joining worker nodes, and optionally deploying the dashboard and flannel network plugin.
This tutorial explains how to integrate CephFS storage into Kubernetes by detailing both static and dynamic PersistentVolume configurations, providing YAML examples, command-line steps, and troubleshooting tips for successful multi‑pod read‑write access.
The article analyzes a reported Kubernetes bug where watch events return non‑monotonic resourceVersion values after deleting early objects, details experiments that failed to reproduce it in version 1.6, explains the underlying code change in 1.7+, and outlines the community fixes introduced in PR #58547.
This guide explains how to configure both static and dynamic PersistentVolumes for CephFS in Kubernetes, covering PV and PVC definitions, deployment integration, the cephfs‑provisioner StorageClass, and common pitfalls such as path and permission issues.
This article explains how Kubernetes health checks—Readiness and Liveness probes—work, the differences between them, practical scenarios for their use, and how to configure probe types, delays, and thresholds to build more resilient cloud‑native applications.
An overview of Kubernetes introduces its role as a container orchestration engine, highlights self‑healing and zero‑downtime deployment, and explains key components such as kubectl, master and worker nodes, pods, deployments, services, secrets, and how they work together in a cloud‑native environment.
This article details the log‑collection difficulties inherent to containerized workloads on the 360 HULK cloud platform and explains how a custom log‑controller, together with Logstash, Qconf, and Qbus2.0, provides dynamic, loss‑free, and source‑aware logging for Kubernetes clusters.
This article explains how to configure Kubernetes Horizontal Pod Autoscaler (HPA) to scale workloads based on custom business metrics collected by Prometheus, covering installation of Metrics Server, deployment of a demo app, setup of the Prometheus adapter, and practical load‑testing steps.
This article explains why microservice architectures need reliable communication, load balancing, fault tolerance, monitoring, tracing and circuit breaking, and shows how Istio—a cloud‑native service mesh built on Envoy—provides these capabilities, enabling blue‑green and canary deployments, traffic routing, retries, and observability within Kubernetes.
This guide explains how to collect and visualize Kubernetes performance metrics—CPU, memory, network, and filesystem—by deploying Heapster with a StatsD sink that forwards data to Splunk's Metrics Store, then using Splunk SPL commands and dashboards for real‑time monitoring.
Facing rapid traffic growth and MySQL limitations, Tongcheng migrated its ticketing order database to a sharded MySQL cluster combined with TiDB, achieving transparent sharding, real‑time synchronization, high‑availability monitoring, and seamless handling of billions of rows and peak QPS over 10,000.
This article explains Kubernetes storage use cases, introduces PersistentVolume and PersistentVolumeClaim concepts, and provides step‑by‑step instructions for creating both static and dynamic Ceph RBD volumes—including secret creation, PV/YAML definitions, StorageClass setup, PVC binding, pod deployment, verification commands, and a discussion of Ceph RBD limitations.
At the 9th Global Operations Conference, Huawei Cloud’s chief architect Cai Xiaogang presented a three‑pronged AIOps strategy that combines large‑scale Kubernetes management, causal tracing in Serverless environments, multi‑source RCA analysis, and clustering‑based black‑box network packet inspection, showcasing how academia‑industry collaboration accelerates cloud‑native operations.
This tutorial walks through installing net‑tools on CentOS, disabling firewalld, setting up Docker CE, pulling and tagging required Kubernetes images, configuring kubelet and kubeadm, initializing a master node, joining worker nodes, installing Flannel networking, and troubleshooting common kubectl issues.
This article from the 360 Qihu operations team explains Kubernetes' architecture, core components, and the three‑step registration‑watch‑action process, while also detailing custom load‑balancer and log‑collection controllers they built for HULK deployments, offering insights for developers and operators.
This article explores why modern enterprises need high‑availability container clouds, outlines the challenges of scaling and stability, and details practical implementations using Kubernetes, HA‑Proxy, OpenResty, and Zuul to achieve resilient, low‑cost, and flexible service delivery across multiple regions.
This article surveys the architecture of cluster resource managers and task schedulers—covering definitions, design principles, and three main categories (centralized, two‑level, and shared‑state) with concrete examples such as Hadoop YARN, Mesos, Spark Drizzle, Borg and Kubernetes—while highlighting their trade‑offs in scalability, fault‑tolerance, and flexibility.
This article explores how Kubernetes has emerged as the dominant container‑orchestration platform, evolving into a low‑level, multi‑cloud standard that creates network effects for distributed tools, reshapes vendor offerings, drives new business models such as Helm‑based package management and serverless functions, and hints at a future where Kubernetes is as invisible as the OS kernel.
This step‑by‑step guide shows how to configure Route53 DNS, prepare a VM with required tools, create an S3 state store, provision a Kubernetes cluster on AWS using kops, validate it, expose a sample service, and clean up the resources.
Tencent Cloud adopted Kubernetes in 2016 for its Borg‑derived design, extensive ecosystem, strong community, and stateful workload support, building a cloud‑native PaaS with managed Kubernetes, integrated monitoring, logging, Helm‑based configuration, high‑availability control‑plane containers, and a suite of container services, while planning broader runtimes and DevOps tools.
This article summarizes the key principles from Red Hat's whitepaper on container‑based application design, explaining classic software design rules and seven Red Hat cloud‑native principles that help build reliable, observable, immutable, disposable, self‑contained, and resource‑constrained containers for Kubernetes.
This article summarizes Red Hat's cloud‑native container design principles, explains each rule—from single concern to runtime confinement—shows how they extend classic software design concepts, and provides practical best‑practice tips and reference links for building resilient, observable, and immutable containerized applications on Kubernetes.
This article recounts Didi’s Elastic Cloud journey, explaining why a private cloud was needed, why containers and Kubernetes were chosen, detailing the overall architecture, product features, networking, monitoring, logging, storage solutions, and sharing practical insights and future plans.
This article explains Kubernetes’ fundamental design principles—including declarative state, level‑trigger messaging, and the list‑watch mechanism—by walking through a ReplicationController‑based runtime flow, detailing each component’s role, and analyzing the reliability, ordering, and real‑time requirements of the watch system.
This article walks you through deploying Flannel (host‑gw and VXLAN modes) and Calico (IP‑IP mode, RR deployment, and BGP configuration) on a Kubernetes cluster using systemd units, etcd backend settings, and detailed CNI configuration files, complete with troubleshooting tips and code snippets.
With large organizations moving to self‑organized microservice teams, this article explains how service mesh and microgateway technologies provide decentralized, consistent service‑to‑service communication, detailing gateway functions, data‑plane and control‑plane roles, and the evolution of low‑invasive components, reverse proxies, sidecar patterns, and native infrastructure such as Istio.
This article explains how Kubernetes NetworkPolicy provides namespace‑ and pod‑level network isolation, how Calico translates policies into iptables rules, and walks through configuration templates, test setups, runtime details, packet‑flow analysis, and practical considerations for production use.
The article explains how Kubernetes has won the container‑orchestration battle, is evolving into a universal low‑level layer for multi‑cloud deployments, and is reshaping software standards, package management, serverless workloads, and the economics of distributed systems.
Learn how to build a repeatable, secure Spinnaker pipeline that automatically runs database migrations and deploys a Go‑based API service to Kubernetes, using Docker images, Run Job stages, and environment variables, without writing custom scripts.
This article explains what Helm Charts are, why they are essential for Kubernetes deployments, how to install and use them—including creating, upgrading, rolling back, and deleting releases—and how JFrog Artifactory enables a private, enterprise‑grade Helm repository.
The 2017 KubeCon+CloudNativeCon in Austin showcased over 4,000 attendees, highlighted Kubernetes as the "Linux of the cloud," explored multi‑cloud strategies, service mesh adoption, storage challenges on Kubernetes, and underscored China's growing influence in the cloud‑native ecosystem.
JD’s Moon Landing Machine Learning Platform, built on Docker and Kubernetes, showcases a cloud‑native architecture that integrates AI services, multi‑tenant security, GPU management, big‑data scheduling, and advanced networking and storage solutions for high‑performance inference and training workloads.
This article explains the Kubelet service lifecycle, its core functions, code structure, startup sequence, and the detailed steps Kubernetes follows to create and run Pods, providing code snippets and diagrams for developers who want to understand or extend Kubelet internals.
Google Cloud’s engineers explain how microservice architectures benefit from small, fast-deployable containers, using Docker, Docker‑Compose, and Kubernetes features such as Pods, Replication Controllers, Services, Labels, and canary releases to achieve automated, high‑availability deployments at Google’s massive scale.
This article compares Linux containers and virtual machines, explaining their architectural differences, use‑case scenarios, security considerations, and how to choose the appropriate technology for applications ranging from single services to full LAMP stacks.
Alibaba has open‑sourced its lightweight container runtime Pouch under Apache 2.0, detailing its evolution from early LXC experiments to a million‑scale deployment, highlighting strong isolation, rich “rich container” features, P2P image distribution, and its integration with Kubernetes and other cloud‑native ecosystems.
This guide details practical solutions for TDH Community Edition operational problems, covering component configuration paths, log locations, disk‑space alerts, service start/stop via kubectl, time synchronization, network/firewall checks, hostname resolution, and license service management.
Istio, an open-source service mesh created by IBM, Google, and Lyft, offers a uniform platform for integrating microservices, managing traffic, enforcing policies, and collecting telemetry, with a control plane that abstracts underlying cluster platforms like Kubernetes, illustrated through developer presentations and slide decks.
This article compares Linux containers and virtual machines, explaining their differing purposes, resource usage, security implications, typical use cases, and how to choose the right technology based on workload requirements and tool ecosystems such as Docker, Kubernetes, VirtualBox, and VMware.
In this interview, JD.com’s infrastructure director Bao Yongcheng discusses his work on large‑scale container platforms, kernel‑level scheduling improvements, power‑saving experiments, and the future of cloud‑native data centers, highlighting practical lessons from the 11.11 shopping festival preparation.
Tencent Cloud Container Service has joined the inaugural CNCF Certified Kubernetes Conformance Program, confirming its Kubernetes platform meets the foundation’s consistency, portability and reliability standards, and giving enterprises such as Wall Street Snippets and Xiaohongshu confidence that workloads will run seamlessly across any certified Kubernetes environment.
This article explains how Cloud TiDB leverages Kubernetes and tidb‑operator to achieve automated operation management, dynamic scaling, resource isolation, and high availability for a cloud‑native distributed database.
This article explains how Cloud TiDB, a cloud‑native NewSQL database launched by UCloud and PingCAP, leverages Kubernetes integration, operators, and custom volume management to provide elastic scaling, high availability, and efficient lifecycle management for distributed workloads.
Rancher is an open‑source software platform that simplifies running and managing Docker and Kubernetes by providing integrated infrastructure services such as networking, storage, load balancing, DNS, security, multi‑cloud deployment, container orchestration, an application catalog, and enterprise‑grade access control.
The article explains how to combine Distributed TensorFlow with Kubernetes—using GlusterFS storage, Deployments for parameter servers, Jobs for workers, service discovery, monitoring, and a Jinja2‑generated YAML template—to create isolated, scalable training clusters with Jupyter and TensorBoard access.
This article compiles a curated collection of high‑quality resources—including introductory explanations, in‑depth tutorials, Chinese translations, and tool overviews for Istio, Linkerd, and Envoy—to guide readers from basic concepts to advanced practices in service mesh and cloud‑native microservice architectures.
At the 2017 Hangzhou Cloud Expo, Alibaba announced the open‑source release of its internally‑developed container runtime Pouch, detailing its seven‑year evolution, unique security and scalability features, collaborations with Zhejiang University, and the company’s strategy to foster an ecosystem‑driven, standards‑focused container technology future.
This article analyzes the technical and organizational factors that determine when Docker containers are appropriate versus virtual machines, covering scenarios from small startups to large enterprises, and presents ten deployment modes ranging from public‑cloud VMs to advanced Kubernetes and hybrid solutions.
This article explains Kubernetes Service types, introduces Ingress and the Traefik reverse‑proxy, and demonstrates a practical high‑availability setup that combines LVS with Traefik to expose services externally, including detailed deployment, Service, and Ingress configurations.
Since September 2016, JD’s AI foundation platform has leveraged Docker and Kubernetes to build a scalable machine‑learning infrastructure that now runs over 5,000 container instances, supports more than 20 AI services, and provides unified GPU, storage, networking, and multi‑tenant capabilities for both inference and training workloads.
