The article reviews traditional availability metrics such as Success‑Ratio, Error‑Budget, MTTR/MTTF, SLA/SLO, and highlights their limitations, then introduces Google’s User‑Uptime and Windowed User‑Uptime metrics, explains their definitions, challenges, experimental results, and why they provide a more user‑centric view of service reliability.
The platform centralizes message‑center management for e‑commerce by adding end‑to‑end tracing, real‑time metrics, and unified logging, enabling business users to query message links, view dashboards, automate retries and approvals, dramatically reducing manual monitoring, improving completion rates above 90%, and paving the way for cost‑optimized, data‑driven operations.
The article outlines JD's data‑driven business development strategy, describing the current challenges of its business data marketplace, the governance framework—including layered architecture, standardization, ClickHouse dictionary refresh, and optimization measures—and the resulting performance improvements and future outlook.
The paper presents a two‑stage e‑commerce search framework where the coarse‑ranking stage is redesigned with multi‑objective optimization, expanded negative sampling, and listwise distillation—guided by a new global transaction hitrate metric—enabling it to surpass fine‑ranking on large candidate sets and boost overall GMV by about one percent.
This article explains how to design effective Prometheus metrics, choose appropriate labels and buckets, and leverage Grafana visualizations to monitor online services, offline processing, and batch jobs, offering concrete examples and best‑practice recommendations for reliable application observability.
This article examines how fragmented metrics and local optimizations can harm overall software quality, discussing the bug‑vs‑feature debate, various measurement approaches, and the importance of viewing quality as a collective team capability rather than an individual or departmental responsibility.
This article summarizes four years of agile transformation in the JD Cloud Platform R&D department, detailing cultural building, continuous agile practices, practical guidelines, effectiveness measurement, and insight analysis across five key areas to improve organizational, technical, and project agility.
The article explains Kubernetes’ built‑in Horizontal Pod Autoscaler, then details the custom Water Pod Autoscaler (WPA) that extends HPA with dual‑signal (load and SOA registration) detection, dual‑threshold scaling, noise filtering, configurable cooldown, frequency limits, tolerance buffers, and integrated alerting for reliable elastic scaling.
This article explains how to design a high‑availability SMS system by identifying stability bottlenecks, defining reliability goals, implementing failover strategies for Redis, MySQL and external services, establishing a comprehensive observability framework, and measuring key quality metrics to ensure 99.99% uptime.
The article details ByteDance’s DevMind platform, describing its origins, the challenges of measuring software development efficiency, the collaborative and value‑driving “flywheel” concepts, the architectural design across data lifecycle and query engine layers, and the principles and future roadmap for scaling R&D performance.
This article explains how to design effective Prometheus metrics, choose appropriate vectors, labels, buckets, and naming conventions, and offers Grafana usage tricks to help engineers monitor online services, batch jobs, and offline processing systems with clear, actionable insights.
The article explains the concepts, classifications, construction methods, and practical usage of data tags and metrics in the financial sector, illustrating how to build indicator and label systems and how to apply them effectively for refined customer operations and business management.
Kubernetes 1.26, themed “Electrifying,” introduces 37 enhancements—including registry changes, storage upgrades, signed release artifacts, Windows high‑privilege containers, metric and scheduling improvements—while promoting 11 features to stable, deprecating 12 APIs, and emphasizing sustainability and carbon‑footprint awareness.
This article explores the three pillars of cloud‑native observability—metrics, logging, and tracing—detailing their definitions, relationships, and practical implementation with tools like Prometheus, ELK/EFK, and SkyWalking, while offering guidance on metric design, collection, visualization, and alerting.
Prometheus is an open‑source monitoring solution that covers metric exposition, scraping, storage, querying, visualization, and alerting, and this guide walks through its architecture, configuration, custom exporters, PromQL queries, Grafana integration, and alert management, providing a comprehensive introduction for developers and ops engineers.
This article explains how to use the MySQL innodb_metrics table to transform the unreadable output of the show engine innodb status command into clear, queryable metrics, including buffer pool statistics and CPU counters, and shows how to enable or disable specific counters as needed.
This guide explains how to collect essential Windows metrics with Prometheus, set up Grafana dashboards for CPU, memory, disk, network, and process monitoring, and configure alert rules, while also comparing self‑hosted and Alibaba Cloud Prometheus solutions for seamless Windows observability.
The article outlines a complete data‑quality‑management framework that first captures business expectations, then translates them into basic and personalized measurement rules, defines four assurance approaches for handling violations, and scales operation with indicators, tooling, and metrics to continuously improve data quality across the lifecycle.
This article explains how RocketMQ's message lifecycle can be fully observed using OpenTelemetry‑based metrics, covering producer, broker, and consumer stages, and shows practical monitoring, alerting, and troubleshooting practices for cloud‑native deployments.
This article explains how to use the Prometheus HTTP API to query instant and range metrics, interpret the JSON responses, and fetch data programmatically with Python, providing code examples and details on request parameters, error handling, and practical usage.
The article examines why engineers favor reuse, outlines the logical arguments that link reuse to productivity, exposes common misconceptions and hidden costs, and proposes autonomy and consistency metrics as practical criteria for deciding when and how to abstract shared components in software systems.
This guide explains what performance testing is, why it’s essential, the key metrics such as throughput, response time, and bandwidth, outlines a step‑by‑step testing process, compares load, stress, endurance and capacity testing types, and reviews popular tools like JMeter, LoadRunner and NeoLoad.
This article visually explains Prometheus’s architecture, core features, metric collection methods, exporters, PromQL query language, and alerting workflow, helping readers understand how to monitor cloud‑native systems effectively while noting its strengths and limitations.
This article explains how RocketMQ integrates with Prometheus and Grafana to provide comprehensive metrics, tracing, and logging, detailing the exporter architecture, deployment choices, span topology, dashboard examples, and ARMS‑based alerting for cloud‑native message‑queue observability.
This guide details key Redis monitoring metrics—including performance, memory, activity, persistence, and error indicators—along with practical commands, configuration settings, and code snippets to help operators efficiently track and troubleshoot Redis instances.
This article visually explains Prometheus by covering its architecture, key features, metric collection methods, exporter role, PromQL query language, and alerting mechanisms, helping readers understand how to monitor cloud‑native systems effectively in production.
This article explains how Spring Boot 3.0.0‑RC1 integrates Micrometer Observation API to provide unified metrics, logging, and distributed tracing, showing the observation lifecycle, configuration steps, sample server and client code, Docker‑compose setup, and notes on native image support for comprehensive application observability.
Programming productivity, also known as software or development productivity, examines how output relates to input, covering definitions, measurement models such as COCOMO II and Jones’s factors, function points, value‑based engineering, and human aspects, while discussing efficiency, effectiveness, profitability, and various factors influencing individual and team efficiency.
Learn how to use Prometheus's stable V1 API to query instant and range data, handle responses and errors, and fetch metrics programmatically with curl and Python, enabling data analysis, cost management, and advanced monitoring beyond basic alerts.
This article explains why measuring a design system is essential, outlines how Baidu's Light Design System built a PATS metric framework through research and four practical steps, and shares real-world results that improved usability, reliability, and overall workflow efficiency.
This article details the redesign of an internal monitoring system, explaining why the original zzmonitor fell short, how Prometheus and its ecosystem were selected, the architecture that integrates remote storage with M3DB, performance benchmarks, Grafana visualisation, and a custom alerting solution.
This article explains how to assess hardware resources, calculate required Elasticsearch cluster size based on data volume, and perform indexing and search benchmark tests to ensure stable performance and optimal throughput for log and metric workloads in production environments.
This article walks through adding complete observability to Spring Boot 3 applications using Micrometer's Observation API, covering metrics, tracing, log correlation, configuration, code examples for both server and client, and even native image support for production-ready monitoring.
This article explains how Spring Boot 3.0.0‑RC1 integrates Micrometer Observation to provide unified metrics, logging, and distributed tracing, showing the required dependencies, configuration, code examples for both server and client, and how to visualize data with Prometheus, Grafana, Loki, and Tempo.
This article outlines the essential components of operational monitoring, covering monitoring objectives, methods, core processes, key tools, metrics for hardware, system, application, network, and business layers, as well as alerting, handling, and best practices for building a comprehensive, reliable monitoring solution.
Ctrip’s internal Dashboard monitoring platform, originally built on HBase, was redesigned by migrating its core writer and storage components to a hybrid VictoriaMetrics‑ClickHouse solution, delivering faster queries, higher write stability, and full Prometheus compatibility while keeping the user experience unchanged.
This article introduces the Prometheus monitoring platform, explains Alertmanager's grouping, inhibition and silencing features, describes Grafana's visualization and alerting capabilities, and provides a detailed course syllabus covering installation, configuration, and advanced monitoring techniques across various environments.
This article explains how to monitor and analyze front‑end performance using the deprecated performance.timing API and the modern PerformanceObserver, detailing key web‑vital metrics such as TTFB, FCP, LCP, FID, and CLS, with code examples and practical interpretation guidelines.
This article outlines the DevMind system—a comprehensive, data‑driven framework that turns R&D efficiency measurement into an online, low‑threshold, scalable practice, covering four best‑practice pillars, technical architectures, product modules, and real‑world impact within large organizations.
This article presents DeepFlow, an open‑source, highly automated observability platform that uses eBPF to provide zero‑code AutoMetrics and AutoTracing, integrates with Prometheus, OpenTelemetry and SkyWalking, and enables SRE, DevOps and NewOps teams to build full‑stack metrics and blind‑spot‑free tracing for cloud‑native applications.
This article details the evolution of NetEase's self‑built time‑series database EyesTSDB into a cloud‑native, second‑level monitoring solution, covering its architecture, core features, integration with VictoriaMetrics, custom plugin workflow, CMDB linkage, real‑world use cases, and future challenges.
This article details the design, data‑collection pipeline, monitoring stack, visualization practices, and alert‑response workflow of a globally deployed Alibaba Cloud‑native system that uses ACK, Prometheus, Grafana, and ARMS to achieve end‑to‑end observability across metrics, tracing, and logs.
This article explains how to enable JMX for Kafka, retrieve and interpret key metrics such as UnderReplicatedPartitions, and troubleshoot common issues like broker failures, disk outages, and replica lag by examining metric values and related logs.
This article details Ctrip's internal Dashboard monitoring platform, explains why its HBase‑based TSDB became a bottleneck, and describes the step‑by‑step migration to a hybrid VictoriaMetrics‑ClickHouse solution with upgraded writers, unified query APIs, performance gains, and future roadmap.
This article explains the fundamental assumptions of linear regression, compares linear and nonlinear models, discusses multicollinearity, outliers, regularization, heteroscedasticity, VIF, stepwise regression, and reviews essential evaluation metrics such as MAE, MSE, RMSE, R² and Adjusted R².
This article details the redesign of Ctrip's internal Dashboard monitoring system, explaining why the original HBase‑based TSDB was replaced, the new architecture using VictoriaMetrics and ClickHouse, component upgrades, unified query layer, performance gains, and future roadmap for metrics handling.
This article explains why measuring is essential for continuous delivery, introduces four valuable metrics—deployable package count, cycle time, mean time between failures, and mean time to recovery—and offers practical tips to improve delivery speed and reliability.
This article explains how to collect comprehensive MySQL performance metrics—including connections, buffer cache, locks, statement counts, throughput, server variables, and slow‑query analysis—using only MySQL's native SHOW commands, providing a fast, low‑overhead monitoring solution.
The article outlines Bilibili’s refined SLO framework—categorizing services into four business tiers, selecting availability, latency, and freshness SLIs, setting concrete SLO targets, and employing multi‑window error‑budget and consumption‑rate alerting strategies to improve stability and provide comprehensive quality dashboards.
Learn how to set up Prometheus to scrape Nexus repository metrics, configure authentication, and create insightful Grafana dashboards that visualize component, Jetty, and JVM metrics, enabling proactive troubleshooting and resource optimization for Nexus services.
This article outlines Snowball's data warehouse background, compares major modeling approaches such as ER, dimensional, DataVault and Anchor models, describes the current challenges of their dimensional model, and details the OneData methodology—including OneModel, OneID, and OneService—along with its practical implementation, results, and future plans.
This tutorial explains how to install Aeraki Mesh sample applications, forward ports to Prometheus and Grafana, and use these tools to query and visualize L7 protocol metrics for Dubbo and Thrift services within an Istio service mesh.
This article explains how massive online game user data can be turned into insightful visualizations by distinguishing in‑game and out‑of‑game contexts, focusing on core metrics such as activity, retention, and payment, and recommending suitable chart types and web‑based visualization tools.
The article explains the importance of measuring software architecture, outlines the granularity of metrics from code to infrastructure, and provides concrete measurement indicators for code implementation, component design, architecture design, and runtime infrastructure to guide effective architecture governance.
The team implements a sustainable “discover‑analyze‑validate” loop that uses a performance blacklist, user‑centric metrics such as interactive time and second‑open rate, automated SDK data collection, and continuous trend monitoring to pinpoint front‑end bottlenecks, apply targeted fixes, and verify measurable load‑time improvements.
This article details a practical approach to constructing a semi‑automated monitoring solution for small‑scale services using Prometheus, covering active monitoring concepts, metric types, service‑framework integration, Grafana dashboards, Alertmanager routing, and deployment on Mesos.
This comprehensive guide walks you through Prometheus fundamentals, including metric exposure, scraping, storage, querying with PromQL, custom exporter creation in Go, dynamic configuration reloading, and visualizing data with Grafana, while also covering alerting with Alertmanager and best practices for accurate histogram bucket design.
This article provides a detailed overview of Prometheus, covering its architecture, metric exposure, scraping models, storage format, metric types, custom exporter implementation in Go, PromQL query language, built‑in functions, Grafana integration, and alerting with Alertmanager, offering practical code examples throughout.
This article explains the core H5 page performance metrics such as FCP, LCP, CLS, TTI, and TBT, describes how to measure them with Lighthouse, and provides comprehensive network‑ and rendering‑layer optimization techniques along with real‑world case studies from the PLUS team to improve user experience.
This article provides a thorough overview of server performance testing, covering fundamental concepts, the influence on users and revenue, key metrics, various test types, a step‑by‑step testing workflow, and a comparative analysis of popular performance testing tools.
This comprehensive guide defines growth, distinguishes it from development, explains how to assess and sustain growth, outlines systematic methodologies, experimentation, growth models, organizational capabilities, and strategic leverage to achieve lasting, scalable product success.
This article describes the design and implementation of an automated FlinkSQL tuning system that monitors metrics, evaluates task health with rule‑based logic, calculates optimal resource adjustments, and performs fast scaling to reduce cluster waste, lower operational costs, and maintain SLA compliance.
After a three‑day, ten‑person investigation into a mini‑program image‑upload failure, we designed and implemented an end‑to‑end observability platform using MDD and SRE principles, defining SLI/SLO, instrumenting client, network, gateway and backend layers, and visualizing metrics with Grafana, ClickHouse and Prometheus.
This article provides a step‑by‑step tutorial on installing and configuring Spring Boot Admin, including Maven dependencies, server and client setup, YML properties, security, Nacos registration, email notifications, custom health indicators, and Micrometer metrics to monitor Spring Boot services.
Architects should treat system capacity like a pre‑flight checklist, using this comprehensive guide to monitor resource usage across services, databases, and queues, and to define business metrics and state‑machine indicators that reveal bottlenecks and guide scaling decisions.
This article explores the factors influencing video quality, explains why continuous improvement is essential, examines the importance and challenges of both subjective and objective quality assessments, and introduces Lingjing—a comprehensive, standards‑based video quality evaluation platform that addresses confidence issues and supports diverse testing scenarios.
After a year of cross‑BU experimentation, the Tianji precision testing platform was built to centralize coverage data, automate incremental analysis, and provide actionable insights, resulting in higher test coverage, faster release gating, and measurable productivity gains for both developers and QA teams.
Prometheus is an open‑source monitoring system that scrapes metrics from services or exporters, stores them in a time‑series database, lets users query with PromQL, visualizes data via its web UI or Grafana, and sends alerts through Alertmanager, supporting custom Go metrics, various discovery methods, and four metric types.
Bilibili’s SRE team, confronting rapid growth and complex systems, built a systematic stability operation that includes emergency response, incident handling, on‑call scheduling, and an Event Operations Center platform, using metrics like MTTR, MTTI and AI‑assisted automation to reduce downtime and improve reliability.
The Meituan Accommodation data governance team shares how they evolved from ad‑hoc, single‑point fixes to a systematic, automated governance framework—covering management, standards, capability, execution, evaluation, and vision—using standardization, digitization, and systematization to achieve measurable quality, cost and efficiency gains across thousands of data assets.
The article explains how technical leaders can apply quantitative analysis and systematic measurement to optimize both software system performance and organizational workflows, using a picture‑recognition service example and a real‑world DevOps incident to illustrate the need for end‑to‑end process mapping, bottleneck identification, and continuous improvement.
This tutorial walks through the end‑to‑end setup of JMeter performance testing data collection using Backend Listener, sending metrics to InfluxDB, and visualizing real‑time TPS, response time, and error rates in Grafana, as well as monitoring OS metrics with node_exporter, Prometheus, and Grafana.
Bilibili’s SRE team combines stability theory, detailed fault‑stage and operational metrics, and a unified emergency‑response platform—including on‑call scheduling, fault‑command incident commanders, automated fault portraits, and rapid post‑mortems—to transform frequent incidents into data‑driven, collaborative recoveries and lay groundwork for AI‑assisted self‑healing.
In this interview series, three industry experts explain how monitoring differs from observability, the shifts required for ops, developers, and architects, the core methodologies and technologies behind metrics, traces, and logs, and practical guidance for selecting and integrating observability tools in cloud‑native environments.
This article provides a comprehensive guide to Micrometer, covering its core metric types, MeterRegistry usage, tagging conventions, and practical code examples, and shows how to integrate it with Spring Boot, Prometheus, and Grafana for end‑to‑end application monitoring.
This article outlines the end‑to‑end workflow for front‑end monitoring in an APM platform, covering data collection, reporting, cleaning, storage, and consumption, and dives deep into environment info, exception handling, performance metrics, and efficient data upload strategies.
The document outlines the background, vision, current status, technical research, value, product and technical architecture, and functional design of a cloud‑native monitoring platform that integrates SkyWalking and Prometheus to provide comprehensive APM, resource utilization, alerting, and rapid fault localization for business and technical middle‑platform services.
This article explains how to use Prometheus to create a precise, end‑to‑end alerting system that shortens detection and diagnosis time, integrates logs and metrics, routes alerts to the right owners, and prevents overwhelming alert storms in production environments.
This article explains the challenges of long request chains in micro‑service architectures, reviews Google’s Dapper tracing requirements, introduces OpenTracing and OpenCensus standards, compares their strengths, and details how OpenTelemetry unifies tracing, metrics and logs with practical integration steps and best‑practice guidance.
This article explains the monitoring challenges introduced by microservice and container evolution, why Prometheus is the preferred observability solution in the cloud‑native era, and presents a comprehensive, multi‑tenant, high‑availability architecture with practical techniques for data collection, storage, query optimization, security, and future trends.
This article explains how to use Spring Boot Actuator to expose Tomcat performance metrics, retrieve health and metric endpoints via HTTP, configure custom endpoints, and optionally log or export the data to Prometheus for continuous monitoring and analysis.
This article explains how Kubernetes Horizontal Pod Autoscaler (HPA) balances resource demand and workload by automatically scaling pod replicas, describes the different metric types it supports, walks through the core controller code (Run, worker, reconcile, and replica calculation), highlights current limitations, and shares practical observations from real‑world usage.
This article outlines nine key DevOps metrics and KPIs—including the four DORA indicators and five additional measures—explaining how they help teams monitor performance, improve delivery speed, ensure quality, and maintain reliable, high‑availability applications.
The article outlines a three‑level indicator system that quantifies engineering quality across efficiency, quality, stability, and resource dimensions, using high‑level result metrics, detailed level‑2 breakdowns, and actionable level‑3 measures to enable drill‑down analysis, risk‑warning, and continuous, data‑driven improvement.
The article introduces a new R&D efficiency metric—throughput standard deviation—explains its statistical basis, shows how it was derived from annual reports, illustrates its application across multiple teams, and discusses practical insights and limitations for software development operations.
This article explains how to build a comprehensive monitoring system for both infrastructure and applications, introducing the USE (Utilization‑Saturation‑Errors) method, key performance metrics, and practical components such as Prometheus, Grafana, full‑link tracing, and the ELK stack to detect and diagnose performance bottlenecks.
This article explains the most important DevOps metrics—such as deployment frequency, lead time, automated test pass rate, change failure rate, MTTR, and others—and how tracking them helps teams improve software delivery speed, quality, and operational efficiency.
To raise front‑end quality, the team implements gray‑release monitoring that triggers log analysis at a 5 % rollout, automatically generates reports within ten minutes, and uses dynamic thresholds and noise‑reduction tactics to detect errors early, enabling rapid rollback or expansion and markedly improving stability and release efficiency.
Even when individual algorithm modules show higher accuracy or revenue, the overall platform can decline due to factors like competitor encroachment, macro‑economic shifts, concept drift, overlapping marginal returns, attribution errors, and coupled A/B experiments, all of which require careful analysis and mitigation.
This article explains how to design effective Prometheus metrics for various application types, choose appropriate vectors, labels, and buckets, and offers Grafana tricks for visualizing dimensions and linking tooltips, providing a comprehensive guide for robust observability.
This article outlines practical Prometheus monitoring techniques, covering how to choose metrics, define labels, select vectors and buckets, and use Grafana tips to build reliable observability for various application types.
This guide explains what vmagent is, its key features, and provides step‑by‑step instructions to install, configure, and verify vmagent on a Kubernetes cluster, including namespace and RBAC setup, custom scrape configs, monitoring endpoints, and troubleshooting tips.
This article outlines practical guidelines for designing Prometheus metrics, covering how to define monitoring targets, choose appropriate vectors and labels, name metrics and labels correctly, select histogram buckets, and leverage Grafana features to visualize and troubleshoot data effectively.
The article explains why and how to use the GSM (Goal‑Signal‑Metric) model in product design, detailing its definition, applicable scenarios, step‑by‑step implementation, and a real‑world case study that demonstrates its value for aligning business, user, and design objectives.
This article compares Kubernetes pod monitoring metrics with traditional KVM/VM metrics across CPU, memory, disk, and network, explaining the underlying reasons for the differences and offering guidance on interpreting the data for effective performance troubleshooting.
This article explains three frequent mistakes in churn analysis—misinterpreting churn rates, falling into Simpson's paradox, and incorrectly inferring causality—illustrated with game‑related examples and emphasizes the need to combine multiple metrics for accurate conclusions.
This article introduces the essential concepts of data quality management, covering the six key quality dimensions, detailed rule and metric templates, data profiling techniques, a systematic quality assurance workflow, and practical data cleaning methods to improve overall data governance.
This article explains how Sentry Relay logs to stderr, configures error reporting, provides health‑check endpoints, and exports a comprehensive set of metrics—including counters, timers, and gauges—to a StatsD server for enterprise‑grade data security monitoring.
This article details the configuration file location, command‑line overrides, and all available Relay settings—including mode, networking, caching, size limits, logging, metrics, and internal error reporting—providing a comprehensive reference for deploying Sentry Relay in production environments.