Calibre monitors web performance through three data sources — Synthetic Testing, Real User Monitoring, and CrUX — all organised under Sites within your Organisation. This page explains the key concepts you’ll encounter as you set up monitoring, interpret results, and integrate Calibre into your workflow.
Core concepts#
The building blocks of your Calibre account and the fundamentals of web performance measurement.
Organisation#
Your top-level account, containing all Teams, Sites, people, and billing settings. Administrators manage organisation-wide settings, invite people, and control access. Learn more about Organisation settings.
Team#
Groups people within an Organisation and controls which Sites they can access. Members can only see Sites belonging to their Team (unless they’re an administrator). Any plan above Starter includes unlimited Teams, so you can organise work however suits your organisation. Learn more about Teams.
Site#
Represents a website or web application you’re monitoring. Each Site belongs to a Team and is the container for all three data sources: Synthetic Testing, Real User Monitoring, and CrUX. Sites appear on the Your Sites overview alongside their Core Web Vitals from each data source.
Site Slug#
The unique, URL-safe identifier for a Site, used when interacting with Calibre’s CLI, APIs, and integrations. Found in Site Settings. Learn more about managing Sites via automation.
Test Pack#
Adds extra Synthetic Testing capacity to any Calibre plan. Test Packs increase the number of tests available per month, allowing teams to monitor more Sites, Pages, or Test Profiles without changing their plan tier. Learn more about Test Packs.
Core Web Vitals#
The three metrics Google uses as search ranking signals: Largest Contentful Paint (LCP) for loading performance, Interaction to Next Paint (INP) for responsiveness, and Cumulative Layout Shift (CLS) for visual stability. Calibre tracks Core Web Vitals across all three data sources. Learn more about analysing Core Web Vitals.
Field Data#
Performance measurements from real users visiting your site. In Calibre, field data comes from two sources: Real User Monitoring (RUM) and the Chrome User Experience Report (CrUX). Field data reflects actual user experience across diverse devices, networks, and conditions.
Lab Data#
Performance measurements from a controlled, simulated environment. In Calibre, lab data comes from Synthetic Testing, where pages are tested using consistent device, network, and location settings. Lab data is reproducible and useful for catching regressions before they reach users.
Percentile (P50, P75, P95)#
Indicates the value below which a given percentage of measurements fall. P75 (75th percentile) is the most commonly used in Calibre — it means 75% of measurements were at or below that value. Core Web Vitals thresholds are evaluated at the 75th percentile.
Page Weight#
The total transfer size of all resources downloaded during a page load. Page weight directly affects load time, especially on slower networks. Calibre tracks it over time and breaks it down by resource type (JavaScript, CSS, images, fonts, and others).
Asset Count#
Counts every HTTP request made during a page load — documents, scripts, stylesheets, images, fonts, and other resources. A high asset count can increase load times, particularly on slower connections.
Performance Score#
Summarises overall page performance as a Lighthouse score from 0 to 100, based on a weighted combination of key metrics. Refer to the Metrics Guide for details on each metric.
Accessibility Score#
Evaluates how accessible a page is to users with disabilities, reported as a Lighthouse score from 0 to 100. Covers colour contrast, ARIA attributes, keyboard navigation, and semantic HTML.
SEO Score#
Checks whether a page follows search engine optimisation best practices, reported as a Lighthouse score from 0 to 100. Covers meta tags, link text, crawlability, and mobile-friendliness.
Best Practices Score#
Checks whether a page follows modern web development best practices, reported as a Lighthouse score from 0 to 100. Covers HTTPS usage, correct image aspect ratios, and avoidance of deprecated APIs.
Real User Monitoring (RUM)#
Concepts specific to measuring performance from your actual visitors.
Real User Monitoring#
Collects performance data from your actual visitors via a lightweight tracking snippet. RUM provides real-time insight into how users experience your site across all browsers and devices, including metrics like Good UX Sessions, live visitor counts, and audience breakdowns by location, device, and browser. Data can be filtered by Page Grouping, navigation type, and other dimensions. Learn more about RUM.
Session#
Tracks a single user visit, capturing performance metrics for that visitor’s experience, including Core Web Vitals and navigation details. Calibre’s RUM is privacy-first: sessions don’t use cookies, don’t collect personally identifiable information, and can’t be linked across visits.
Sampling Rate#
Controls what percentage of user sessions RUM collects data from. A rate of 100% captures every visitor, while a lower rate (such as 15%) reduces data volume while still providing statistically representative results. Configured in Site → Real User Monitoring → Settings. Learn more about installing RUM.
Page Grouping#
Groups related URLs together so you can see performance metrics for types of pages — such as all blog posts or product pages — rather than individual URLs. Groupings can be used as filters across all RUM reports and in the Pages Leaderboard. Learn more about Page Groupings.
Good UX Sessions#
Shows the percentage of RUM sessions where all three Core Web Vitals were rated "good.", alongside other UX metrics. This single number summarises overall user experience quality.
Chrome User Experience Report (CrUX)#
Concepts specific to Google’s public dataset of real-user performance metrics.
Chrome User Experience Report#
Google’s public dataset of real-user performance metrics from Chrome users who have opted in, covering millions of websites. CrUX measures Core Web Vitals and other user experience signals. Calibre integrates CrUX data directly into its dashboards, so you can track field metrics without setting up BigQuery or additional tooling. Learn more about CrUX in Calibre.
Experience Distribution#
Visualises what percentage of your audience has a good, needs-improvement, or poor experience for a given metric. Available in CrUX and RUM reports, helping you understand how performance varies across your user base.
Metric Subparts#
Breaks a metric into its contributing components so you can pinpoint exactly where time is being spent. For example, LCP splits into Time to First Byte, Image Load Delay, Image Load Duration, and Image Render Delay. Subparts are available in CrUX and RUM reports.
Synthetic Testing#
Concepts specific to automated, lab-based performance testing.
Synthetic Testing#
Runs automated performance tests using Google Lighthouse and a recent version of Chrome on Calibre’s Test Agents. Synthetic tests use controlled conditions — specific devices, network speeds, and locations — so results are consistent and comparable over time. This is Calibre’s lab data source. Learn more about Synthetic Testing.
Page#
A URL you want to monitor with Synthetic Testing. Each Page is tested during every Snapshot using all of the Site’s Test Profiles — for example, a homepage, a product listing page, and a checkout page would each be a separate Page. Learn more about Pages.
Test Profile#
Defines how a synthetic test runs — which device, connection speed and other parameters like Third Party Blocking. Examples include "Mobile on a 4G connection from Sydney" or "Desktop from London." Each Site can have multiple Test Profiles, and every Snapshot runs tests across all Pages using every Test Profile. Test Profiles can also include custom cookies, HTTP headers, and third-party blocking rules. Learn more about Test Profiles.
Snapshot#
A single run of all synthetic tests for a Site, producing metrics and audits for every Page and Test Profile combination. Snapshots can run automatically on a schedule, be triggered manually, or be created via the API in response to a deployment. Learn more about Snapshots.
Pulse#
Shows how a single Page performs over time. Pulse displays Web Vitals for the most recent Snapshot across all Test Profiles, with 7- or 30-day metric history, Deploy Markers, and Performance Budget status. Learn more about Pulse.
Performance Budget#
Sets a threshold for a metric — when it’s exceeded, Calibre alerts you via Slack, email, or webhook. Each budget targets a specific Page and Test Profile combination and has three states: Met, At Risk, or Exceeded. Learn more about Performance Budgets.
Single Page Test#
Runs a one-off performance test against any public URL, independent of Sites and ongoing monitoring. Useful for quick checks, comparing pages, sharing results with colleagues, or running tests as part of a CI/CD pipeline. Each test produces a full Lighthouse Report with metrics, audits, a filmstrip, and a HAR file. Learn more about Single Page Tests.
Deploy Marker#
Marks when a code deployment occurred on performance charts. Deploy Markers correlate changes in performance with specific releases, making it easier to identify which deployment caused a regression or improvement. They can be created via the CLI, Node.js API, or GraphQL API, and typically include a revision identifier such as a commit SHA. Learn more about deploy tracking.
Test Agent#
Executes Calibre’s synthetic tests by running Google Lighthouse with a recent version of Chrome on dedicated hardware. Each test starts with a fresh browser profile (no cache or cookies). Agents are available across 17 global locations and can use dynamic or fixed IP addresses.
Test Location#
Where a Test Agent physically runs synthetic tests. Choosing a location close to your users or servers produces more realistic TTFB and latency measurements. Calibre offers locations across North America, Europe, Asia Pacific, and more.
Device Emulation#
Simulates a specific device’s screen size, pixel density, and user agent string during a synthetic test. Calibre offers a range of device presets (including mobile phones, tablets, and desktops) that can be configured in a Test Profile.
Network Throttling#
Simulates a specific network connection speed during a synthetic test. Calibre supports connection profiles from 2G through to cable, controlling bandwidth and latency to replicate real-world network conditions. Configured as part of a Test Profile.
Test Schedule#
Controls how often Snapshots run automatically for a Site. Options include Off, Daily (at a specific hour), Hourly, or Every X Hours. Adjustable when adding a Site or later in Site settings. Learn more about Test Schedules.
Test Mode#
Determines which audits and data are collected during a Snapshot. Standard Mode runs everything. Reduced Audits omits Accessibility, SEO, and Best Practices scores. Minimum Audits further omits the Long Task Timeline and Third Party data. Lower modes can help tests complete when pages have issues. Learn more about Test Modes.
Test Verification#
Ensures stable and reliable test results by analysing variations in key metrics. When a result deviates significantly from the baseline, it is flagged and re-tested, and the most reliable result is released. Learn more about Test Verification.
Lighthouse Report#
The full output of a Google Lighthouse audit, produced during each synthetic test. Includes scores for Performance, Accessibility, SEO, and Best Practices, along with specific audits grouped into Opportunities, Diagnostics, and Passed Audits. Learn more about Lighthouse audits in Calibre.
Lighthouse Audit#
Evaluates a specific aspect of a page within a Lighthouse Report, such as "Serve images in next-gen formats" or "Ensure text remains visible during webfont load." Audits are grouped into Opportunities (actionable improvements with estimated savings), Diagnostics (additional information), and Passed Audits. Learn more about audits.
Filmstrip#
Captures screenshots at intervals during a page load, showing the visual progress of rendering over time. Helps identify when content becomes visible and when visual changes stop. Available in Snapshot and Single Page Test results.
Form Authentication#
Lets you test pages that require login. Calibre fills in a login form with credentials you provide, supporting single and multi-step forms. After authentication completes, the browser cache is cleared so the actual test starts from a clean state. Configured per-Site in Synthetic → Settings → Authentication. Learn more about testing logged-in pages.
HAR File#
Records every network request made during a page load — URLs, response codes, timings, headers, and transfer sizes — in the standard HTTP Archive format. HAR files can be downloaded from Snapshot and Single Page Test results for detailed analysis.
Third Party#
Any resource loaded from a domain you don’t control — analytics scripts, advertising tags, social media widgets, fonts, and CDN-hosted libraries. Calibre’s third-party report quantifies the performance cost of each external provider, showing transfer size and main thread execution time. Third-party requests can also be blocked in Test Profiles to measure their impact. Learn more about Third Party tracking.
Waterfall Chart#
Displays every network request made during a page load as horizontal bars on a timeline, with each bar showing timing phases (DNS lookup, connection, TLS, waiting, download). Helps identify bottlenecks such as render-blocking resources, slow server responses, or long request chains.
Long Task Timeline#
Shows JavaScript tasks that took longer than 50 milliseconds during a page load — the tasks that block the browser’s main thread and degrade responsiveness. Displays parent and child tasks with source information, and allows filtering by first-party versus third-party scripts. Learn more about the Long Task Timeline.
Render Video#
Records the visual rendering of a page during a synthetic test, showing exactly what a user would see as the page loads. Useful for identifying visual regressions, layout shifts, and loading delays.
Integrations#
Notifications, APIs, and workflow connections.
Insights Report#
Delivers an email digest of notable metric changes for your Sites. Available weekly (sent Mondays) or monthly (sent on the first of the month), highlighting regressions, improvements, and Core Web Vitals assessment changes. Available on all plan tiers. Learn more about email notifications.
Team Insights Report#
Summarises performance across all Sites in a Team in a monthly email. Each Site is rated as Regressed, Improved, or Stable based on changes over the past two months, giving everyone a bird’s-eye view of how sites are tracking. Sent to all Team members by default. Learn more about email notifications.
Budget Alert#
Fires when a Performance Budget changes state — from Met to At Risk, or from At Risk to Exceeded. Alerts can be sent via Slack, email, or webhook.
Slack Integration#
Delivers Budget Alerts, Snapshot summaries, and other notifications directly to a Slack channel. Learn more about the Slack integration.
Webhook#
Sends an HTTP callback when specific events occur — a Snapshot completing, a Performance Budget changing state, and more. Use webhooks to trigger custom workflows, update dashboards, or integrate with tools like Zapier. Payloads are signed with an HMAC signature for verification. Learn more about Webhooks.
Pull Request Review#
Automatically runs performance tests against pull request preview deployments and posts results as a GitHub comment. Helps catch performance regressions before code is merged. Learn more about Pull Request Reviews.
GitHub Actions#
Lets you run Calibre Single Page Tests or trigger Snapshots as part of your GitHub CI/CD workflow. Learn more about the GitHub Actions integration.
CLI#
Manages Sites, Snapshots, Single Page Tests, Deploy Markers, and more from the terminal. Useful for scripting, CI/CD pipelines, and automation workflows. View the CLI command reference.
Node.js API#
Interacts with Calibre programmatically from Node.js applications. Can manage Sites, Snapshots, Test Profiles, Deploy Markers, and more — the recommended approach for custom automation. Learn more about the Node.js API.
GraphQL API#
The underlying API that powers the CLI and Node.js API, available for advanced use cases not covered by other interfaces. Accessed via an API Token over HTTPS with a rate limit of 10 requests per second. Learn more about the GraphQL API.
HTTP Site API#
Creates Snapshots and Deploy Markers via a lightweight JSON-over-HTTP interface with no runtime dependencies — ideal for post-deploy hooks in CI/CD pipelines. Authenticated with a Site Secret rather than an API Token. Learn more about the HTTP Site API.
API Token#
Authenticates requests to Calibre’s CLI, Node.js API, and GraphQL API. Calibre offers two types: Admin Tokens (full access to all Sites) and Personal Access Tokens (scoped to the user’s permissions). Tokens can be set to expire after a chosen period and restricted by IP address. Learn more about API Tokens.
Site Secret#
Authenticates requests to the HTTP Site API for a single Site. Found in Site Settings alongside the Site Slug. Unlike API Tokens, a Site Secret is scoped to one Site and only works with the HTTP Site API.