Upgrade to Pro — share decks privately, control downloads, hide ads and more …

๐Ÿ‡ณ๐Ÿ‡ด NDC Oslo 2022

๐Ÿ‡ณ๐Ÿ‡ด NDC Osloย 2022

โ„น๏ธ Web performance APIs you (probably) didn't know existed

Responsiveness to interaction is crucial for our apps and youโ€™ve probably heard about the amazing tools we have to measure web performance. But did you know there are many performance APIs native to the Web Platform?

In this talk, weโ€™ll see how to leverage the power of some of them to reliably measure responsiveness and correctly identify culprits for bad experiences.

Matheus Albuquerque

September 28, 2022
Tweet

More Decks by Matheus Albuquerque

Other Decks in Technology

Transcript

  1. Hello, NDC Oslo ๐Ÿ™‹ ๐Ÿ‡ณ๐Ÿ‡ด ๐ŸŒ WEB PERFORMANCE APIS YOU

    (PROBABLY) DIDN'T KNOW EXISTED โ€ข THE 28TH OF SEPTEMBER, 2022.
  2. ร… Iโ€™M MATHEUS ๐Ÿ™‹ โ† @YTHECOMBINATOR ON THE WEB โ†

    SR. SOFTWARE ENGINEER @MEDALLIA โ† MENTOR @TECHLABS
  3. WEB PERFORMANCE APIS YOU (PROBABLY) DIDN'T KNOW EXISTED #2 60

    MINUTES IS A LOT OF TIME, I KNOWโ€ฆ ๐Ÿฅฑ
  4. WEB PERFORMANCE APIS YOU (PROBABLY) DIDN'T KNOW EXISTED #3 THE

    IDEA BEHIND THIS SESSION IS HEAVILY INSPIRED BY ZENO'S SESSION BACK IN 2016.
  5. โ† HTTP/2 (REQ/RES MULTIPLEXING; EFFICIENT COMPRESSION; REQUEST PRIORITIZATION; SERVER PUSH)

    โ† COMPRESS JAVASCRIPT (E.G. WITH BROTLI) โ† IMAGE CDNS (40โ€“80% SAVINGS; LESS WORK WITH TRANSFORMATION, OPTIMIZATION, AND DELIVERY) YOU ALREADY KNOWโ€ฆ
  6. โ† OPTIMIZE THIRD-PARTIES (DEFER/REPLACE/ UPDATE THEM; TRY EXPERIMENTAL IDEAS) โ†

    PRPL PATTERN (PUSH CRITICAL RESOURCES โ†’ RENDER THE INITIAL ROUTE ASAP โ†’ PRE-CACHE IN BACKGROUND โ†’ LAZILY LOAD OTHER STUFF) โ† WORKERS TO OFF-LOAD THE MAIN THREAD YOU ALREADY KNOWโ€ฆ
  7. This talk presents Why am I here? Measuring Measuring (with

    the platform) Optimizing (with the platform) Why are we here? Closing thoughts
  8. This talk presents Why am I here? Measuring Measuring (with

    the platform) Optimizing (with the platform) Why are we here? Closing thoughts
  9. #QUESTION ๐Ÿค” What does it mean to be fast? If

    you were to summarize web performance in one metric, whatโ€™d be your pick?
  10. #QUESTION ๐Ÿค” What does it mean to be fast? If

    you were to summarize web performance in one metric, whatโ€™d be your pick? e.g. load time, responsiveness, etc.
  11. WEB PERFORMANCE APIS YOU (PROBABLY) DIDN'T KNOW EXISTED If you

    were to summarize web performance in one metric, whatโ€™d be your pick?
  12. WEB PERFORMANCE APIS YOU (PROBABLY) DIDN'T KNOW EXISTED If you

    were to summarize web performance in one metric, whatโ€™d be your pick? 1st/2nd/3rd ANSWERS โ† REACT BRUSSELS DISCOUNT CODE ๐Ÿ‡ง๐Ÿ‡ช OTHERS โ† STICKERS ๐Ÿ’…
  13. #PROTIP๐Ÿ’ก A lot of loading metrics don't capture user experience.

    We need to think about our metrics in terms of what matters.
  14. #QUESTION ๐Ÿค” What does it mean to be fast? If

    you were to summarize web performance in one metric, whatโ€™d be your pick?
  15. WEB PERFORMANCE APIS YOU (PROBABLY) DIDN'T KNOW EXISTED If you

    were to summarize web performance in one metric, whatโ€™d be your pick?
  16. MEASURING PERCEIVED LOAD SPEED HOW QUICKLY A PAGE CAN LOAD

    AND RENDER ALL OF ITS VISUAL ELEMENTS TO THE SCREEN LOAD RESPONSIVENESS HOW QUICKLY A PAGE CAN LOAD/RUN ANY REQUIRED JS IN ORDER FOR COMPONENTS TO RESPOND TO USER INTERACTION RUNTIME RESPONSIVENESS AFTER THE PAGE LOAD, HOW QUICKLY CAN THE PAGE RESPOND TO USER INTERACTION? SMOOTHNESS DO TRANSITIONS & ANIMATIONS RENDER AT A CONSISTENT FRAME RATE AND FLOW FLUIDLY?
  17. #PROTIP๐Ÿ’ก The only way to truly know how your site

    performs for your users is to actually measure its performance as those users are loading and interacting with it.
  18. โ† THE FIRST EVER RUM EXPERIMENT IN 1746 โ† HE

    GATHERED หœ200 MONKS IN CIRCUMFERENCE, WIRED USING IRON, AND DISCHARGED A BATTERY THROUGH THE HUMAN CHAIN โ† HE WAS TESTING THE LATENCY OF AN ELECTRIC SIGNAL WITH REAL USERS JEAN-ANTOINE NOLLET
  19. #PROTIP๐Ÿ’ก Browsers now provide us with many apis to fetch

    performance metrics that help site owners make sites faster.
  20. MEASUREMENT APIS GENERAL STUFF DOMHighResTimeStamp, PerformanceObserver, PAGE VISIBILITY API &

    LAYOUT INSTABILITY API PROFILING LONG TASKS API, UserAgentSpecificMemory & JS SELF- PROFILING API TIMING USER TIMING, ELEMENT TIMING. EVENT TIMING, RESOURCE TIMING, NAVIGATION TIMING & SERVER TIMING SENSORS BATTERY STATUS & NETWORK INFORMATION
  21. GENERAL STUFF DOMHighResTimeStamp MICROSECOND RESOLUTION & BASED ON THE TIME

    AT WHICH THE USER NAVIGATED TO THE PAGE LAYOUT INSTABILITY ALL LAYOUT SHIFT EVENTS ON A PAGE PerformanceObserver SUBSCRIBE TO PERFORMANCE-RELATED EVENTS PAGE VISIBILITY KNOW WHEN A DOCUMENT BECOMES VISIBLE OR HIDDEN
  22. โ† CANCEL NETWORK REQUESTS AND OTHER TIME- BASED EVENTS โ†

    SUSPEND WEB SOCKETS AND EVENTSOURCE CONNECTIONS โ† PAUSE VIDEOS AND CAROUSELS, IF THE USER IS NOT WATCHING THEM PAGE VISIBILITY
  23. โ† FETCH INCOMPLETE MEDIA TO GIVE A COMPLETELY LOADED SITE

    WHEN ACTIVE AGAIN โ† DOWNLOAD THE REST OF YOUR BUNDLES USING DYNAMIC IMPORTS (AND CACHE THEIR ASSETS) โ† MONITOR THE USERโ€™S BEHAVIOR โ† THROTTLE TIMERS LIKE setTimeout() PAGE VISIBILITY
  24. โ† SUBSCRIBE TO PERFORMANCE-RELATED EVENTS โ† CALLBACKS ARE GENERALLY FIRED

    DURING IDLE PERIODS โ† YOU TELL THE OBSERVER WHAT TYPES OF ENTRIES TO LISTEN FOR PerformanceObserver
  25. const observer = new PerformanceObserver((list) = > { for (const

    entry of list.getEntries()) { console.log(entry.toJSON()); } }); observer.observe({type: 'some-entry-type'}); PerformanceObserver
  26. โ† MANY WEBSITES HAVE DOM ELEMENTS SHIFTING AROUND DUE TO

    CONTENT LOADING ASYNCHRONOUSLY โ† THIS API GIVES US A LIST OF ALL LAYOUT SHIFT EVENTS ON A PAGE LAYOUT INSTABILITY
  27. const observer = new PerformanceObserver((list) = > { const entries

    = list.getEntries(); / / Shifts that are not preceded by input events const nonInputEntries = entries.filter((entry) = > !entry.hadRecentInput); for (const entry of nonInputEntries) { console.log(entry); } }); observer.observe({type: 'layout-shift'}); LAYOUT INSTABILITY
  28. [ { "name": "", "entryType": "layout-shift", "startTime": 3087.2349999990547, "duration": 0,

    "value": 0.3422101449275362, "hadRecentInput": false, "lastInputTime": 0 } ] LAYOUT INSTABILITY
  29. TIMING (1st ACT) USER TIMING ALLOWS YOU TO MARK POINTS

    IN TIME AND THEN MEASURE THE DURATION BETWEEN THOSE MARKS. EVENT TIMING EVENT PROCESSING TIME + TIME UNTIL THE NEXT FRAME CAN BE RENDERED. THE BASIS FOR THE FID METRIC. ELEMENT TIMING MEASURE THE RENDER TIME OF SPECIFIC ELEMENTS. THE BASIS FOR THE LCP METRIC.
  30. / / Record the time i m m ediately before

    running a task performance.mark('myTask:start'); await doMyTask(); / / Record the time i m m ediately after running a task performance.mark('myTask:end'); / / Measure the delta between the start and end of the task performance.measure('myTask', 'myTask:start', โ€˜myTask:end'); observer.observe({type: 'measure'}); USER TIMING
  31. EXAMPLE CUSTOM METRICS โ† STYLESHEETS DONE BLOCKING โ† SCRIPTS DONE

    BLOCKING โ† FONTS LOADED โ† TEXT DISPLAYED USER TIMING
  32. โ† ALLOWS YOU TO MEASURE THE RENDER TIME OF SPECIFIC

    ELEMENTS โ† USEFUL FOR KNOWING WHEN THE LARGEST IMAGE OR TEXT BLOCK WAS PAINTED TO THE SCREEN โ† THE BASIS FOR THE LARGEST CONTENTFUL PAINT (LCP) METRIC ELEMENT TIMING
  33. { "name": "http://localhost:3030/public/hero.png", "id": "hero-image", "entryType": "element", "startTime": 90.68500006105751, "duration":

    0, "identifier": "some-identifier", "naturalHeight": 73, "naturalWidth": 70, "intersectionRect": { "x": 253, "y": 30, "width": 71, "height": 73 } } ELEMENT TIMING
  34. โ† USEFUL TO MEASURE THE EVENT PROCESSING TIME AS WELL

    AS THE TIME UNTIL THE NEXT FRAME CAN BE RENDERED โ† EXPOSES A NUMBER OF TIMESTAMPS IN THE EVENT LIFECYCLE โ† THE BASIS FOR THE FIRST INPUT DELAY METRIC EVENT TIMING
  35. const observer = new PerformanceObserver((entryList) = > { const firstInput

    = entryList.getEntries()[0]; const firstInputDelay = firstInput.processingStart - firstInput.startTime; / / Measure the time it takes to run all event handlers const firstInputProcessingTime = firstInput.processingEnd - firstInput.processingStart; / / Measure the entire duration of the event const firstInputDuration = firstInput.duration; }); observer.observe({type: 'first-input'}); EVENT TIMING
  36. MEASURE THE LATENCYโ€ฆ โ† FROM A CLICK UNTIL WE REORDER

    CONTENT ON A TABLE โ† TO DRAG A SLIDER TO FILTER SOME DATA โ† FOR A FLYOUT TO APPEAR WHEN HOVERING A MENU ITEM EVENT TIMING
  37. PROFILING LONG TASKS API REPORTS TASKS THAT TAKES LONGER THAN

    50 MS AND IT'S THE BASIS FOR TTI AND TBT METRICS JS SELF-PROFILING API PROFILE SPECIFIC COMPLEX OPERATIONS AND IDENTIFY HOT SPOTS USING A SAMPLING PROFILER UserAgentSpecificMemory DETECT MEMORY LEAKS IN APPS THAT HANDLE A HUGE VOLUME OF DATA
  38. โ† REPORTS TASKS THAT TAKES LONGER THAN 50 MS โ†

    USEFUL TO TRACK WHEN THE BROWSER'S MAIN THREAD IS BLOCKED โ† THE BASIS FOR TIME TO INTERACTIVE (TTI) AND TOTAL BLOCKING TIME (TBT) METRICS LONG TASKS
  39. observer.observe({type: 'longtask'}); { name: "same-origin-descendant", entryType: "longtask", startTime: 1023.40999995591, duration:

    187.19000002602115, attribution: [ { name: "unknown", entryType: "taskattribution", startTime: 0, duration: 0, containerType: "iframe", containerSrc: "child.html", containerId: "", containerName: "child1" } ] }; LONG TASKS
  40. const profiler = new Profiler({ sampleInterval: 10, maxBufferSize: 10000 });

    / / Do work . . . const trace = await profiler.stop(); sendProfile(trace); JS SELF-PROFILING
  41. โ† PROFILE SPECIFIC COMPLEX OPERATIONS โ† PROFILE THIRD-PARTY SCRIPTS โ†

    PROFILE THE ENTIRE PAGE LOAD PROCESS โ† COMBINE WITH OTHER EVENTS/IMPORTANT METRICS, LIKE LONG TASKS OR EVENT TIMING APIS JS SELF-PROFILING
  42. โ† MEASURES THE MEMORY USAGE AND DETECTS MEMORY LEAKS โ†

    WAITS FOR THE NEXT GC AND THEN MEASURES MEMORY IMMEDIATELY AFTER THE UNNEEDED MEMORY HAS BEEN RELEASED โ† FORCES GC IF IT DOES NOT HAPPEN FOR 20S UserAgentSpecificMemory
  43. โ† FORGETTING TO UNREGISTER AN EVENT LISTENER โ† ACCIDENTALLY CAPTURING

    OBJECTS FROM AN IFRAME โ† NOT CLOSING A WORKER โ† ACCUMULATING OBJECTS IN ARRAYS MEMORY LEAKS
  44. const obj = { a: new Array(1000), b: new Array(2000)

    }; setInterval(() = > { console.log(obj.a); }, 1000); MEMORY LEAKS
  45. โ† APPS THAT HANDLE A HUGE VOLUME OF DATA โ†

    REGRESSION DETECTION DURING THE ROLLOUT OF A NEW VERSION โ† A/B TESTING TO EVALUATE MEMORY IMPACT โ† CORRELATING MEMORY USAGE WITH USER METRICS TO UNDERSTAND THE IMPACT OF MEMORY USAGE UserAgentSpecificMemory
  46. { name: "same-origin-descendant", entryType: "longtask", startTime: 1023.40999995591, duration: 187.19000002602115, attribution:

    [ { name: "unknown", entryType: "taskattribution", startTime: 0, duration: 0, containerType: "iframe", containerSrc: "child.html", containerId: "", containerName: "child1" } ] } { bytes: 1000000, breakdown: [ { bytes: 1000000, attribution: [ { url: "https://example.com", scope: "Window", }, ], types: ["JS", "DOM"], }, { bytes: 0, attribution: [], types: [], }, ], } { "frames": [ { "name": "Profiler" }, { "column": 0, "line": 100, "name": "", "resourceId": 0 }, { "name": "set innerHTML" }, { "column": 10, "line": 10, "name": "A", "resourceId": 1 } { "column": 20, "line": 20, "name": "B", "resourceId": 1 } ], "resources": [ "https://example.com/page", "https://example.com/app.js", ], "samples": [ { "stackId": 0, "timestamp": 161.99500000476837 }, { "stackId": 2, "timestamp": 182.43499994277954 }, { "timestamp": 197.43499994277954 }, { "timestamp": 213.32999992370605 }, { "stackId": 3, "timestamp": 228.59999990463257 }, ], "stacks": [ { "frameId": 0 }, { "frameId": 2 }, { "frameId": 3 }, { "frameId": 4, "parentId": 2 } ] } LONG TASKS SELF-PROFILING USERAGENT MEMORY ๐Ÿคฏ
  47. TIMING (2nd ACT) RESOURCE TIMING RETRIEVE AND ANALYZE DETAILED NETWORK

    TIMING DATA REGARDING THE LOADING SERVER TIMING COMMUNICATE PERFORMANCE METRICS ABOUT HOW TIME IS SPENT WHILE PROCESSING THE REQUEST NAVIGATION TIMING COMPLETE TIMING INFORMATION FOR NAVIGATION OF A DOCUMENT
  48. const observer = new PerformanceObserver((list) = > { for (const

    entry of list.getEntries()) { / / If transferSize is 0, the resource was fulfilled via the cache. console.log(entry.name, entry.transferSize === 0); } }); observer.observe({type: 'resource'}); RESOURCE TIMING
  49. const observer = new PerformanceObserver((list) = > { for (const

    entry of list.getEntries()) { console.log('Time to first byte', entry.responseStart); } }); observer.observe({type: 'navigation'}); NAVIGATION TIMING
  50. / / ServiceWorker startup time const workerStartupTime = entry.responseStart -

    entry.workerStart; / / Request time only (excluding redirects, DNS, and connection/TLS time) const requestTime = entry.responseStart - entry.requestStart; / / Response time only (download) const responseTime = entry.responseEnd - entry.responseStart; / / Request + response time const requestResponseTime = entry.responseEnd - entry.requestStart; NAVIGATION TIMING
  51. MEASURE ANY WORK THAT THE SERVER DOES TO COMPLETE A

    REQUEST: โ† ROUTING/AUTHENTICATING THE REQUEST โ† QUERYING DATABASES/API CALLS TO THIRD-PARTY SERVICES โ† RUNNING CONTENT THROUGH TEMPLATING SYSTEMS SERVER TIMING
  52. SENSORS BATTERY STATUS INFORMATION ABOUT THE POWER SOURCE, CHARGE LEVEL,

    AND OTHERS NETWORK INFORMATION ADAPT THE USERSโ€™ EXPERIENCE BASED ON THE QUALITY OF THEIR CONNECTION
  53. โ† SWITCH BETWEEN SERVING HIGH/LOW DEFINITION CONTENT BASED ON THE

    USER'S NETWORK โ† DECIDE WHETHER TO PRELOAD RESOURCES โ† DEFER UPLOADS/DOWNLOADS WHEN USERS ARE ON A SLOW CONNECTION โ† ADAPT TO SITUATIONS WHEN USERS ARE OFFLINE NETWORK INFORMATION
  54. โ† ENABLE OFFLINE MODE IF THE NETWORK QUALITY IS NOT

    GOOD ENOUGH โ† WARN USERS THAT DOING SOMETHING OVER CELLULAR COULD COST THEM MONEY โ† USE IT IN YOUR ANALYTICS TO GATHER DATA ON YOUR USERS' NETWORK QUALITY NETWORK INFORMATION
  55. switch (connectionType) { case "4g": return <Video src={videoSrc} />; case

    "3g": return <Image src={imageSrc.hiRes} alt={alt} />; default: return <Image src={imageSrc.lowRes} alt={alt} />; } NETWORK INFORMATION
  56. โ† MANY PERFORMANCE OPTIMIZATIONS CAN BE MADE WHEN WE CAN

    PREDICT WHAT USERS MIGHT DO โ† RESOURCE HINTS ARE A SIMPLE BUT EFFECTIVE WAY TO ALLOW DEVELOPERS TO HELP THE BROWSER TO STAY ONE STEP AHEAD OF THE USER AND KEEP PAGES FAST RESOURCE HINTS
  57. < ! - - Preconnect - - > <link rel="preconnect"

    href="https://fonts.gstatic.com"> <link rel="preconnect" href="https://scripts.example.com"> < ! - - Preloading - - > <link rel="preload" href="https://example.com/fonts/font.woff"> < ! - - DNS Prefetch - - > <link rel="dns-prefetch" href="https://fonts.gstatic.com"> <link rel="dns-prefetch" href="https://images.example.com"> < ! - - Prefetch - - > <link rel="prefetch" href="/uploads/images/pic.png"> <link rel="prefetch" href="https://example.com/news/?page=2"> < ! - - Prerender - - > <link rel="prerender" href="https://example.com/news/?page=2"> RESOURCE HINTS
  58. < ! - - Preconnect - - > <link rel="preconnect"

    href="https://fonts.gstatic.com"> <link rel="preconnect" href="https://scripts.example.com"> < ! - - Preloading - - > <link rel="preload" href="https://example.com/fonts/font.woff"> < ! - - DNS Prefetch - - > <link rel="dns-prefetch" href="https://fonts.gstatic.com"> <link rel="dns-prefetch" href="https://images.example.com"> < ! - - Prefetch - - > <link rel="prefetch" href="/uploads/images/pic.png"> <link rel="prefetch" href="https://example.com/news/?page=2"> < ! - - Prerender - - > <link rel="prerender" href="https://example.com/news/?page=2"> RESOURCE HINTS
  59. โ† IT ALLOWS THE BROWSER TO SETUP EARLY CONNECTIONS BEFORE

    THE REQUEST IS ACTUALLY SENT TO THE SERVER. THIS INCLUDES: โ€ข TLS NEGOTIATIONS โ€ข TCP HANDSHAKES โ† ELIMINATES ROUNDTRIP LATENCY PRECONNECT
  60. 100MS 200MS 300MS 400MS 500MS 600MS 700MS HTML CSS FONT

    1 FONT 2 FONTS START LOADING FONTS RENDERED
  61. 100MS 200MS 300MS 400MS 500MS 600MS 700MS HTML CSS FONT

    1 FONT 2 FONTS START LOADING FONTS RENDERED FONT 1 FONT 2
  62. โ† IT WARNS THE BROWSER ABOUT THE DOMAINS ITโ€™S GOING

    TO NEED TO LOOK UP โ† THIS ENABLES THE IT TO DO MORE WORK IN PARALLEL, REDUCING THE OVERALL LOAD TIME โ† USE IT WHEN YOUR PAGE USES RESOURCES FROM A DIFFERENT DOMAIN DNS PREFETCH
  63. โ† FETCH RESOURCES IN THE BACKGROUND (IDLE TIME) AND STORE

    THEM IN CACHE โ† USE IT WHEN YOU HAVE A REASONABLE AMOUNT OF CERTAINTY โ† THIS SHOULD BE WEIGHED AGAINST THE RISK OF WASTING RESOURCES PREFETCH
  64. โ† IT GOES ONE STEP FURTHER AND EXECUTES THE FILES

    โ† IT DOES PRETTY MUCH ALL THE WORK REQUIRED TO DISPLAY THE PAGE, MAKING ITS LOAD FEEL INSTANTANEOUS โ† THE GAMBLE IS EVEN GREATER HERE PRERENDER
  65. โ† INCREASE THE PRIORITY OF THE LCP IMAGE โ† LOWER

    THE PRIORITY OF ABOVE-THE-FOLD IMAGES AND PRELOADED RESOURCES โ† LOWER THE PRIORITY FOR NON-CRITICAL DATA FETCHES โ† REPRIORITIZE SCRIPTS PRIORITY HINTS
  66. < ! - - Increase the priority of the LCP

    image - - > <img src="image.jpg" fetchpriority="high" /> < ! - - Lower the priority of above-the-fold images - - > <ul class="carousel"> <img src="img/carousel-1.jpg" fetchpriority="high" /> <img src="img/carousel-2.jpg" fetchpriority="low" /> <img src="img/carousel-3.jpg" fetchpriority="low" /> </ul> < ! - - Reprioritize scripts - - > <script src="async_but_important.js" async fetchpriority="high"></script> <script src="blocking_but_unimportant.js" fetchpriority="low"></script> PRIORITY HINTS
  67. < ! - - Increase the priority of the LCP

    image - - > <img src="image.jpg" fetchpriority="high" /> < ! - - Lower the priority of above-the-fold images - - > <ul class="carousel"> <img src="img/carousel-1.jpg" fetchpriority="high" /> <img src="img/carousel-2.jpg" fetchpriority="low" /> <img src="img/carousel-3.jpg" fetchpriority="low" /> </ul> < ! - - Reprioritize scripts - - > <script src="async_but_important.js" async fetchpriority="high"></script> <script src="blocking_but_unimportant.js" fetchpriority="low"></script> PRIORITY HINTS
  68. / / Important validation data const user = await fetch("/user");

    / / Less important content data const relatedPosts = await fetch("/posts/suggested", { priority: "low" }); PRIORITY HINTS
  69. / / Important validation data const user = await fetch("/user");

    / / Less important content data const relatedPosts = await fetch("/posts/suggested", { priority: "low" }); PRIORITY HINTS
  70. โ† IMAGES FOR DESKTOP AND TABLET CAN BE หœ2-4X LARGER

    THAN THE MOBILE ONES. โ† WITH srcset, THE BROWSER WON'T DOWNLOAD THE LARGER IMAGES UNLESS THEY'RE NEEDED. THAT SAVES BANDWIDTH. RESPONSIVE IMAGES
  71. <img src="ndc_oslo_logo.png" srcset=" ndc_oslo_logo-300.png 300w, ndc_oslo_logo-600.png 600w, ndc_oslo_logo-1200.png 1200w "

    /> <img src="ndc_oslo_logo.png" srcset=" ndc_oslo_logo-1x.png 1x, ndc_oslo_logo-2x.png 2x, ndc_oslo_logo-3x.png 3x " /> RESPONSIVE IMAGES
  72. <img src="ndc_oslo_logo.png" srcset=" ndc_oslo_logo-300.png 300w, ndc_oslo_logo-600.png 600w, ndc_oslo_logo-1200.png 1200w "

    /> <img src="ndc_oslo_logo.png" srcset=" ndc_oslo_logo-1x.png 1x, ndc_oslo_logo-2x.png 2x, ndc_oslo_logo-3x.png 3x " /> RESPONSIVE IMAGES
  73. <img src="ndc_oslo_logo.png" srcset=" ndc_oslo_logo-300.png 300w, ndc_oslo_logo-600.png 600w, ndc_oslo_logo-1200.png 1200w "

    /> <img src="ndc_oslo_logo.png" srcset=" ndc_oslo_logo-1x.png 1x, ndc_oslo_logo-2x.png 2x, ndc_oslo_logo-3x.png 3x " /> RESPONSIVE IMAGES
  74. #PROTIP ๐Ÿ’ก Lazy-loading iframes can lead to 2-3% median data

    savings, 1-2% FCP reductions, and 2% FID improvements at the 95th percentile. โ€”โ€‰Chrome teamโ€™s research, 2019
  75. SCHEDULING & THE WEB WE HAVE A FEW SCHEDULING PRIMITIVES:

    โ† setTimeout โ† requestAnimationFrame โ† requestIdleCallback โ† postMessage
  76. โ† EVERYONE SHOULD USE THE SAME SCHEDULER โ† HAVING MORE

    THAN ONE SCHEDULER CAUSES RESOURCE FIGHTING โ† WE HAVE TASKS INTERLEAVING WITH BROWSER WORK (RENDERING, GARBAGE COLLECTION) SCHEDULING & THE WEB
  77. NATIVE SCHEDULER โ† A MORE ROBUST SOLUTION FOR SCHEDULING TASKS

    โ† INTEGRATED DIRECTLY INTO THE EVENT LOOP โ† CONTROL AND SCHEDULE PRIORITIZED TASKS IN A UNITED AND FLEXIBLE WAY โ† ALIGNED WITH THE WORK OF THE REACT TEAM AND IN COOPERATION WITH GOOGLE, W3C AND OTHERS
  78. scheduler.postTask() SCHEDULE AND CONTROL PRIORITIZING TASKS. scheduler.wait() YIELD AND RESUME

    AFTER SOME AMOUNT OF TIME OR PERHAPS AFTER AN EVENT HAS OCCURRED. scheduler.yield() BREAK UP LONG TASKS BY YIELDING TO THE BROWSER AND CONTINUING AFTER BEING RESCHEDULED. isInputPending() DETERMINE IF THE CURRENT TASK IS BLOCKING INPUT EVENTS. NATIVE SCHEDULER
  79. scheduler.postTask() scheduler.postTask(() = > { console.log('NDC Oslo'); }, { delay:

    10 }); scheduler.postTask(() = > { console.log('NDC Sydney'); }); scheduler.postTask(() = > { console.log('NDC Minnesota'); }); / / 'NDC Sydney' 'NDC Minnesotaโ€™ 'NDC Oslo'
  80. scheduler.postTask() const controller = new TaskController({ priority: "user-blocking" }); const

    signal = controller.signal; console.log(signal.priority); / / 'user-blocking' console.log(signal.aborted); / / 'false' scheduler.postTask(doWork, { signal }); controller.setPriority("background"); controller.abort();
  81. isInputPending() while (workQueue.length > 0) { if (navigator.scheduling.isInputPending()) { /

    / Stop doing work to handle any input event break; } let job = workQueue.shift(); job.execute(); }
  82. isInputPending() while (workQueue.length > 0) { if (navigator.scheduling.isInputPending()) { /

    / Stop doing work to handle any input event break; } let job = workQueue.shift(); job.execute(); }
  83. isInputPending() while (workQueue.length > 0) { if (navigator.scheduling.isInputPending(['mousedown', 'keydown'])) {

    / / Stop doing work to handle a mouse or key event break; } let job = workQueue.shift(); job.execute(); }
  84. scheduler.yield() async function doWork() { while (true) { let hasMoreWork

    = doSomeWork(); if (!hasMoreWork) { return; } await scheduler.yield(); } }
  85. async function doWork() { while (true) { let hasMoreWork =

    doSomeWork(); if (!hasMoreWork) { return; } if (!navigator.scheduling.isInputPending()) { continue; } await scheduler.yield(); } } scheduler.yield()
  86. scheduler.yield() async function findInFiles(query) { for (const file of files)

    { await yieldOrContinue("user-visible"); for (const line of file.lines) { fuzzySearchLine(line, query); } } }
  87. function resourcefulOperation(value: number) { let newValue = String(value); for (let

    i = 0; i < 1000000; i++) { newValue = `${value} + ${i} = ${value + i}`; } return newValue; } function ResourcefulComponent(props: { value: number }) { const { value } = props; const result = resourcefulOperation(value); return <p>{result}</p>; }
  88. function* resourcefulOperation(value: number) { let newValue = String(value); while (true)

    { yield; for (let i = 0; i < 1000000; i++) { newValue = `${value} + ${i} = ${value + i}`; } return newValue; } } const initialValue = 0; const scheduler = new Scheduler(resourcefulOperation, initialValue); function ResourcefulComponent(props: { value: number }) { const { value } = props; const result = scheduler.performUnitOfWork(value); return <p>{result}</p>; }
  89. function* resourcefulOperation(value: number) { let newValue = String(value); while (true)

    { yield; for (let i = 0; i < 1000000; i++) { newValue = `${value} + ${i} = ${value + i}`; } return newValue; } } const initialValue = 0; const scheduler = new Scheduler(resourcefulOperation, initialValue); function ResourcefulComponent(props: { value: number }) { const { value } = props; const result = scheduler.performUnitOfWork(value); return <p>{result}</p>; } PROMOTED TO A GENERATOR YIELDING EXECUTION DOING CONCURRENT TASKS
  90. enum SchedulerState { IDLE = "IDLE", PENDING = "PENDING", DONE

    = "DONE", } class Scheduler<T> { state: SchedulerState; result: T; worker: (data: T) = > Generator; iterator: Generator; constructor(worker: (data: T) = > Generator, initialResult: T) { this.state = SchedulerState.IDLE; this.worker = worker; this.result = initialResult; } performUnitOfWork(data: T) { switch (this.state) { case "IDLE": this.state = SchedulerState.PENDING; this.iterator = this.worker(data); throw Promise.resolve(); case "PENDING": const { value, done } = this.iterator.next(); if (done) { this.result = value; this.state = SchedulerState.DONE; return value; } throw Promise.resolve(); case "DONE": this.state = SchedulerState.IDLE; return this.result; } } }
  91. โ† CSS TRICKS (E.G. font-display: swap AND will-change) โ† IMPORT

    ON VISIBILITY (USING AN IntersectionObserver) โ† SCHEDULING ALTERNATIVES WITH sendBeacon โ† CANVAS + WORKERS = OffscreenCanvas โ€ฆAND MUCH MORE!
  92. โ† AVOID THE INITIAL HTTP TO HTTPS REDIRECTS SO THAT

    PAGES LOAD FASTER USING HTTP STRICT TRANSPORT SECURITY (HSTS) โ† SEND DATA WITHOUT EXTRA ROUND TRIPS WITH QUIC โ€ฆAND MUCH MORE!
  93. #PROTIP ๐Ÿ’ก Start with observability services or libraries like web-vitals.

    Then create your own abstractions on top of the web (e.g. React hooks).
  94. const Video = lazy(() = > import("./Video")); const Preview =

    lazy(() = > import("./Preview")); const networkInfo = useNetworkStatus(); const { With, Switch, Otherwise } = usePatternMatch(networkInfo); MOVE FAST & BREAK NOTHING โ„ข
  95. <Suspense fallback={<div>Loading . . . </div>}> <With unsupported> <NetworkStatus networkInfo="unsupported"

    /> <Video /> </With> <With effectiveConnectionType="2g"> <NetworkStatus networkInfo="2g" /> <Preview /> </With> </Suspense> MOVE FAST & BREAK NOTHING โ„ข
  96. WEB PERFORMANCE APIS YOU (PROBABLY) DIDN'T KNOW EXISTED Which browsers

    are visiting our app? โ€”โ€‰APP DYNAMICS REPORT, LAST THREE MONTHS
  97. #REALITYCHECK ๐Ÿ˜ณ Phone users experience slow First Input Delay on

    7x more websites. โ€”โ€‰Web Almanac By HTTP Archive, 2021
  98. DESKTOP PHONE 0 25 50 75 100 SLOW ( >

    = 250MS) AVERAGE FAST (<50 MS) FIRST INPUT DELAY
  99. THE TIME IT TAKES TO LOAD A PAGE HOW EASY

    IT IS TO FIND WHAT I'M LOOKING FOR HOW WELL THE SITE FITS MY SCREEN HOW SIMPLE THE SITE IS TO USE HOW ATTRACTIVE THE SITE LOOKS โ€”โ€‰SPEED MATTERS, VOL. 3 HOW IMPORTANT IS SPEED?
  100. #RESEARCH ๐Ÿ“š 40% of Brits reported that they had become

    physically violent toward their computers. โ€”โ€‰British Psychology Society, 2009
  101. #RESEARCH ๐Ÿ“š 38% of Americans with computer issues reported that

    they had screamed, yelled, or physically assaulted their computers. โ€”โ€‰crucial.com research, 2013
  102. #RESEARCH ๐Ÿ“š A 500ms delay resulted in up to a

    26% increase in frustration and up to an 8% decrease in engagement. โ€”โ€‰Radware research, 2013
  103. #RESEARCH ๐Ÿ“š Delayed web pages caused a 38% rise in

    mobile users' heart rates โ€” equivalent to the anxiety of watching a horror movie alone. โ€”โ€‰Ericsson ConsumerLab neuro research, 2015
  104. #RESEARCH ๐Ÿ“š 53% of mobile users abandon sites that take

    over 3 seconds to load. โ€”โ€‰DoubleClick, 2020
  105. #PROTIP๐Ÿ’ก People from different parts of the world have a

    different threshold for frustration. โ€”โ€‰UNDERSTANDING EMOTION FOR HAPPY USERS, BY PHILIP TELLIS
  106. โ† LONG TASKS DELAYED TTI โ† AS FIRST-PAGE LONG TASK

    TIME INCREASED, OVERALL CONVERSION RATES DECREASED โ† MOBILE HAD UP TO หœ12X LONGER LONG TASKS โ† OLDER DEVICES COULD BE SPENDING HALF OF THEIR LOAD-TIME ON LONG TASKS BUSINESS OUTCOMES โ€”โ€‰AKAMAI AND CHROME RESEARCH, 2017
  107. โ† THE RANGE OF BROWSERS THAT PEOPLE USE IS USUALLY

    WIDER THAN WE THINK โ† ACCESS TO 4G AND 5G IS STILL (VERY) LIMITED IN SOME PARTS OF THE GLOBE โ† 4G/5G AVAILABILITY AND 4G/5G SPEED ARE DIFFERENT METRICS WHY ARE WE HERE?
  108. โ† ANYTHING THAT SLOWS DOWN OR PREVENTS PEOPLE FROM ACCOMPLISHING

    THEIR GOALS WILL CAUSE FRUSTRATION โ† PEOPLE WILL ABANDON SITES THAT CAUSE FRUSTRATION WHY ARE WE HERE?
  109. WEB PERFORMANCE APIS YOU (PROBABLY) DIDN'T KNOW EXISTED ALWAYS TRY

    TO CORRELATE BUSINESS METRICS WITH PERFORMANCE. #1
  110. WEB PERFORMANCE APIS YOU (PROBABLY) DIDN'T KNOW EXISTED DON'T TAKE

    FAST NETWORKS, CPUS AND RAM FOR GRANTED. #2
  111. WEB PERFORMANCE APIS YOU (PROBABLY) DIDN'T KNOW EXISTED THERE'S NO

    SILVER BULLET. IDENTIFY YOUR CORE METRICS. #6
  112. WEB PERFORMANCE APIS YOU (PROBABLY) DIDN'T KNOW EXISTED "IN GOD

    WE TRUST. ALL OTHERS MUST BRING DATA." โ€” W. EDWARDS DEMING #7