TechPrufer
Get started
Cloudflare
Cloudflare logoCloudflare · Feature trail

Website Performance Timing Metrics

5 posts · Jul 2012 – Sep 2026

What this trail covers

Cloudflare's understanding of website performance metrics has evolved to go beyond basic Time To First Byte (TTFB). This includes detailed analysis of client-side timing measurements from tools like cURL and Chrome, mapping them to network events like DNS lookup, TCP handshake, and TLS setup. The focus is shifting towards measuring the actual user experience, considering factors like page interactivity and resource loading, rather than just the initial byte arrival. Browser Insights provides Real User Monitoring (RUM) data, and the BEACON dataset expands this by providing billions of real-world performance measurements across 10,000 of the largest websites, updated daily in Google BigQuery. BEACON reports Core Web Vitals (LCP, CLS, INP) as full histograms, allowing for percentile analysis. It also includes LCP and INP sub-parts to separate loading and interaction stages, and analyzes the impact of application architecture like single-page applications (SPAs) with soft navigations. Researchers can combine BEACON with other datasets, such as GDP per capita, to explore correlations between economic conditions and web performance. Cloudflare Radar will integrate BEACON data into its Web Performance section, correlating Internet Quality Index (IQI) with website performance.

All Cloudflare trails
  1. How fast is the web? Explore billions of real-user measurements with BEACONThis post introduces the Cloudflare BEACON dataset, an anonymized collection of billions of real-world performance measurements from 10,000 of the largest websites. BEACON expands the RUM Archive standard by providing full histograms for Core Web Vitals (LCP, CLS, INP) and introduces LCP and INP sub-parts to break down loading and interaction stages. It also analyzes the performance differences between hard and soft navigations in SPAs and explores correlations between web performance and economic indicators like GDP per capita. The dataset is updated daily in Google BigQuery and is intended for researchers to explore web performance across browsers, devices, and countries.
  2. How does Cloudflare’s Speed Test really work?This post details how Cloudflare's Speed Test works, focusing on measuring connection quality and realistic usage patterns rather than peak throughput. It explains the use of Workers and anycast for routing, predefined data block sizes, and the measurement of loaded latency, jitter, and packet loss. The post also introduces the Aggregated Internet Measurement (AIM) score and discusses how to interpret results, highlighting that the test measures the combined performance of the user's device, local network, and the path to Cloudflare's network.
  3. Performance measurements… and the people who love themThis post details Cloudflare's journey to better understand and measure performance, particularly latency. It defines latency from a client's perspective and highlights the challenges of server-side-only measurements. The post adopts Catchpoint's definition of TTFB (DNS + Connect + TLS + Wait) and aligns it with browser timing metrics. It emphasizes the limitations of aggregated percentile data (like p99) by illustrating how a significant portion of requests (the ignored 1%) can still experience high latency, even when p99 appears acceptable. The post introduces the need for new tools and techniques to visualize and understand the full spectrum of performance data, moving beyond simple percentile plots to uncover subtle issues.
  4. Introducing Speed Brain: helping web pages load 45% fasterIntroduces Speed Brain, a new feature that uses the Speculation Rules API to prefetch and prerender content based on predicted user navigation. It leverages insights from global request logs and user interaction events (touch/click) to deliver static content before a user navigates to a page. Initial lab testing showed up to a 75% reduction in Largest Contentful Paint (LCP) with an aggressive loading model, and a 45% reduction in LCP on successful prefetches with the default model. Speed Brain is available to all plan types immediately and at no cost, with plans to offer more aggressive speculation models and automatic configuration in the future. It also encourages enabling Real User Measurements (RUM) for improved and customized prefetching rules.
  5. Faster website, more customers: Cloudflare Observatory can help your business growThis post introduces Cloudflare Observatory, a new unified dashboard for website performance monitoring. Observatory integrates Real User Monitoring (RUM) data and Google Lighthouse synthetic testing, offering regional testing capabilities. It provides customized recommendations for performance improvements, often with a single-click enablement of Cloudflare features like Polish and Image Resizing. The post also details the evolution from the previous Speed Tab, highlighting increased test limits and regional availability for different plan types.
  7. Are you measuring what matters? A fresh look at Time To First ByteThis post argues that Time To First Byte (TTFB) is an insufficient metric for evaluating web page load speed and user experience. It explains the ambiguity of TTFB, contrasting it with Time To Last Byte (TTLB) and highlighting that neither fully captures user perception. The post emphasizes the importance of Real User Monitoring (RUM) and Core Web Vitals (Largest Contentful Paint, First Input Delay, Cumulative Layout Shift) as more accurate indicators of performance. It details how TTFB does not account for render-blocking resources, browser parsing, or the benefits of modern protocols like HTTP/2 and HTTP/3, and how Cloudflare features like Zaraz, Rocket Loader, and Mirage improve loading times without necessarily impacting TTFB.
  8. Introducing Timing Insights: new performance metrics via our GraphQL APIIntroduces 'Timing Insights', a new set of performance metrics exposed via the GraphQL Analytics API. These metrics break down Time to First Byte (TTFB) into `edgeTimeToFirstByteMs`, `edgeDnsResponseTimeMs`, and `originResponseDurationMs`. The data is available in the `httpRequestsAdaptiveGroups` dataset and supports average, median, 95th, and 99th percentiles. The post provides examples of querying these metrics to diagnose performance issues, demonstrating how to identify slow URLs and pinpoint DNS lookup time as a contributing factor to high TTFB.
  9. INP. Get ready for the new Core Web VitalThis post introduces Interaction to Next Paint (INP) as a new Core Web Vital that will replace First Input Delay (FID) in March 2024. It explains the differences between FID and INP, highlighting that INP is a more comprehensive measure of interactivity as it considers all user interactions and includes processing and presentation delays. The post analyzes Cloudflare's RUM data to show that INP is generally more challenging to achieve 'Good' scores for than FID, especially on mobile devices. It provides guidance on how to prepare for the change by enabling RUM and optimizing JavaScript tasks, DOM interactions, and third-party scripts, and highlights Cloudflare's free RUM provider and its ability to remove third-party scripts as ways to help customers improve their INP scores.
  10. Discovering what’s slowing down your website with Web AnalyticsIntroduced Web Analytics Vitals Explorer, which integrates Core Web Vitals (LCP, FID, CLS) and paint timings (First Paint, First Contentful Paint) into the Web Analytics product. This feature provides actionable insights by identifying specific page elements negatively impacting these metrics, along with their CSS selectors, URL paths, and detailed measurement data. The JavaScript beacon was updated transparently to collect this new data.
  11. Improving Performance and Search Rankings with Cloudflare for Fun and ProfitThis post introduces Cloudflare's support for Google's Core Web Vitals (LCP, FID, CLS) as a direct factor in Google Search ranking. It highlights how Cloudflare's performance products (caching, Argo Smart Routing, Image Resizing, APO for WordPress, Workers Sites) directly improve these metrics. It also details Cloudflare Browser Insights as a one-click solution for measuring Web Vitals and announces the upcoming free, privacy-aware Cloudflare Web Analytics product for any website.
  12. Start measuring Web Vitals with Browser InsightsThis post introduces the integration of Google's Web Vitals (Largest Contentful Paint, First Input Delay, Cumulative Layout Shift) into Cloudflare's Browser Insights product. It explains the importance of these metrics for measuring user experience (loading, interactivity, visual stability) and how they complement existing navigation timing metrics. The post highlights the benefits of using Cloudflare's RUM for real-time data collection across all browsers (not just Chromium) and its free availability, enabling users to push a button to start measuring these key performance indicators.
  13. Inside the Web Browser’s Performance APIThis post introduces Cloudflare's Browser Insights product, which leverages the browser's Performance API (specifically Navigation Timing) to collect and analyze client-side website performance metrics. It details the Navigation Timing API's attributes (e.g., domainLookupEnd, connectEnd, responseEnd) and explains how Cloudflare injects a JavaScript beacon to collect these metrics from end-users. The post outlines the specific metrics surfaced in the Cloudflare dashboard (DNS, TCP, Request, Response, Processing, Load Event, Total Time) and their corresponding calculations from the Performance API.
  14. Introducing Browser InsightsThis post introduces Browser Insights, a new tool that measures website performance from the perspective of end-users using Real User Monitoring (RUM). It enables one-click activation via the Speed section of the dashboard. When enabled, a JavaScript snippet is added to HTML pages, utilizing the standard Performance API to collect timing information. The tool provides request-level metrics (e.g., TCP connection time, request time) and page-level metrics (e.g., Page Load Time), which can be filtered by dimensions like URL and Country. The post also outlines future plans for Browser Insights, including more metrics, subresource analytics, A/B testing integration, error collection, alerting, and Cloudflare-powered insights.
  15. A Question of TimingThis post details how to use cURL and Chrome's developer tools to measure various timing metrics for website performance, including time_namelookup, time_connect, time_appconnect, time_pretransfer, and time_starttransfer. It explains what each metric represents in the context of an HTTP over TLS 1.2 connection and discusses the limitations of TTFB as a sole performance indicator, emphasizing the need to measure real user experience.
  16. Data-driven development with Cloudflare Mobile SDKIntroduced Cloudflare Mobile SDK for Android and iOS to provide developers with visibility into their app's network utilization. The SDK helps identify top N requests, slow requests, and requests most likely to fail, correlating network experience data with existing engagement metrics. It emphasizes privacy by not collecting persistent identifiers and is offered for free.
  17. Stop worrying about Time To First Byte (TTFB)This post argues that Time To First Byte (TTFB) is a misleading metric for end-user experience. It demonstrates through experiments with a custom Go server and analysis of WebPageTest and Gomez that TTFB often measures the first byte of the HTTP response headers, not the first byte of the page content. It also shows how gzip compression, while improving total download time, can negatively impact TTFB. The post concludes that TTFB is only useful as a trend measured at the server and that Cloudflare focuses on optimizing the real end-user page visibility time.