This topic describes the page loading speed detection and analysis features in frontend monitoring of Application Real-Time Monitoring Service (ARMS).
After your application is connected to the ARMS frontend monitoring feature, you can view the following performance data of the application on the Page Speed page:
- Details of the page loading time
- Waterfall chart of page loading data
- Page performance distribution
- Traces of slow page sessions
- Page loading distribution
- Custom performance metrics that are manually reported
Go to the Page Speed page
- Log on to the ARMS console .
- In the left-side navigation pane, click Browser Monitoring. On the Browser Monitoring page, click the name of the application that you want to view.
- In the left-side navigation pane, choose .
Details of the page loading time
- In the Page Load Time Details section, the data in the curve chart is displayed based
on the average values of the metrics within the specified time range. The average
value reflects the average performance over a period of time. However, this value
is sensitive to extreme values and large fluctuations. For example, if the overall
page loading speed is slow due to weak network connection during an access request,
the average response time is long. You can click the
icon in the upper-right corner to remove extreme values. This way, the impact of
extreme values on the overall performance trend is avoided.
- If the data in the curve chart increases sharply, you can use the Performance Sample Distribution and Slow Page Session Trace(TOP20) sections to identify the issue.
Waterfall chart of page loading data
In the Page Load Waterfall Plot section, the waterfall chart shows the response time in each stage based on the page loading order. Each type of data in the figure is the average value of a specified metric within a specified time range. To optimize performance, we recommend that you implement the related countermeasures on specific stages.
Page performance distribution
The Performance Distribution section shows the distribution of the page performance.
On the Performance Stacked Area Chart tab, a stack line chart is displayed. This chart uses time as the horizontal axis. You can view the distribution of performance metrics at different points in time.
On the Performance Sample Distribution tab, you can view the sample proportion of the page loading time in the specified time range. For example, you can view the number of pages that can be opened within 1 second and the proportion of samples of long-tail users.
Traces of slow page sessions
The Slow Page Session Trace(TOP20) section provides a performance waterfall chart of static resource loading during the page loading process. In this section, you can view the status of page resource loading based on page performance data. This allows you to identify and handle performance bottlenecks at the earliest opportunity. For more information, see Session tracing.
Page loading distribution
A page is loaded on the browser of a user. The loading time is determined by specific factors. These factors include the geographical location, network condition, browser, and carrier. Therefore, the frontend monitoring feature provides the statistics on geographical distribution, terminal distribution, network distribution, and version distribution. This allows you to identify performance bottlenecks.
Description of performance metrics
| Reported field | Description | Formula | Remarks |
|---|---|---|---|
| FMP | The First Meaningful Paint (FMP). | For more information, see Technical solution of FMP. | None |
| FPT | The FPT. | responseEnd - fetchStart | This field indicates the duration from the point in time when a request is initiated to the point in time when the browser begins to parse the bytes of the first batch of HTML documents. |
| TTI | The time to interact (TTI). | domInteractive - fetchStart | This field indicates the point in time when the browser starts to load resources after it resolves all HTML documents and constructs the document object model (DOM). |
| Ready | The time consumed to complete HTML loading, which is the time consumed to construct the DOM. | domContentLoadEventEnd - fetchStart | If a JS script is synchronously executed on the page, the execution time of the JS script can be calculated based on the following formula: Execution time of the JS script = Ready - TTI. |
| Load | The time consumed to completely load the page. | loadEventStart - fetchStart | This field can be calculated based on the following formula: Load = FPT + DOM + (Ready - TTI) + Res. |
| FirstByte | Time to first packet | responseStart - domainLookupStart | None |
| Reported field | Description | Formula | Remarks |
|---|---|---|---|
| DNS | The time consumed for Domain Name System (DNS) query. | domainLookupEnd - domainLookupStart | None |
| TCP | The time consumed for Transmission Control Protocol (TCP) connection. | connectEnd - connectStart | None |
| TTFB | The time to first byte (TTFB), which indicates the time consumed to respond to a request. | responseStart - requestStart | TTFB can be calculated in different ways. For more information about how TTFB is calculated in ARMS, see Google development definition. |
| Trans | The time consumed for data transmission. | responseEnd - responseStart | None |
| DOM | The time consumed for DOM resolution. | domInteractive - responseEnd | None |
| Res | The time consumed for resource loading. | loadEventStart - domContentLoadedEventEnd | This field indicates the time consumed to synchronously load resources on the page. |
| SSL | The time consumed for Secure Sockets Layer (SSL) connection. | connectEnd - secureConnectionStart | This field is valid only when HTTPS is used to transmit data. |