Official tech channel of 360, building the most professional technology aggregation platform for the brand.
This article describes a lightweight, resource‑efficient log collection solution that combines Filebeat agents, optional Logstash aggregation, and Kafka transport, detailing configuration choices, meta‑persistence, back‑pressure mechanisms, monitoring setup, and deployment architecture for reliable at‑least‑once delivery.
This article explains the problem of inaccurate /proc information in container environments, introduces LXCFS as a FUSE‑based solution that maps cgroup data to /proc, and provides step‑by‑step instructions for installing, mounting, and using LXCFS with Docker and Kubernetes.
This article describes 360's Cloud Connect Network solution, covering its background, application scenarios, design, server architecture, traffic forwarding mechanisms, CCSI traffic isolation, performance optimizations with DPDK, and future enhancements for multi‑VPC and IDC interconnectivity.
This comprehensive tutorial walks through preparing a CentOS 7 host, configuring hostnames, disabling swap and firewalls, installing Docker and Kubernetes components with kubeadm, setting up the master and worker nodes, deploying the flannel network plugin, testing with nginx, and installing the Kubernetes dashboard, all with detailed commands and troubleshooting tips.
This article details the design, deployment, configuration, operation, monitoring, and backup of a PostgreSQL high‑availability cluster built on Patroni, Etcd, and LVS at 360, covering hardware layout, software versions, installation steps, parameter tuning, fail‑over testing, and future outlook.
The article analyzes the instability of shared‑type virtual machines, introduces enterprise‑level instances with fixed CPU scheduling and NUMA topology, details the applied technologies such as CPU pinning, PCI‑Passthrough and multi‑queue NICs, and presents extensive sysbench and STREAM benchmark results that demonstrate superior isolation, stability and performance of enterprise instances over shared ones.
This article explains how 360 Commercialization implements cross‑IDC hot‑standby for Kafka using MirrorMaker 2, covering MM2 fundamentals, architecture, internal topics, deployment on Kubernetes, design goals, solution details, challenges such as dynamic configuration and offset reverse‑mapping, and productized risk mitigation.
This article presents a comprehensive log collection solution that evaluates Filebeat versus Logstash, explains Filebeat's meta‑persistence and back‑pressure mechanisms, outlines design goals and architecture layers, and provides detailed configuration examples for both direct Filebeat‑to‑Kafka and Filebeat‑to‑Logstash‑to‑Kafka pipelines.
This article examines several subtle Go programming issues—including unexpected loop variable addresses, the scope nuances of the := operator, proper handling of goroutine concurrency with worker pools, and how struct field ordering affects memory alignment—providing code examples and practical solutions to avoid these pitfalls.
HTTP/2, now widely supported by browsers and major sites, replaces HTTP/1.1 by introducing a binary framing layer, header compression, stream prioritization, server push, and flow control, addressing head-of-line blocking and inefficiencies while maintaining HTTP semantics, with detailed discussion of its evolution, structure, and deployment considerations.
