Overview

Application launch monitoring helps you understand how fast your Android app becomes usable after a user taps the app icon. Use it to identify slow startup times, track performance regressions, and optimize the user’s first impression of your app.

With this feature, you can:

• Measure time to initial display (TTID) and time to full display (TTFD) for cold and warm starts
• View launch performance trends in the RUM Summary and Mobile Performance Dashboard
• Drill into individual launch events to diagnose bottlenecks

How it works

During initialization, the RUM Android SDK creates a view called ApplicationLaunch. This view’s start time matches the start of the Android process. The ApplicationLaunch view includes any logs, actions, and resources created before your first call to startView.

Time to initial display and time to full display

In Android SDK versions 3.5.0+, the time to initial display (TTID) and time to full display (TTFD) are collected during the application launch period.

Each time to initial display and time to full display is categorized by launch type:

• Cold start: The application is launched from scratch. Cold starts happen when the application is launched for the first time since device boot or since the system terminated the process of the application.
• Warm start: The application is launched using a subset of the operations that take place during a cold start. Warm starts result from situations such as a user backing out of the application or re-launching the application. Warm starts can also result from the user launching the Activity when the OS-process already exists, such as the arrival of a silent push or WorkManager job execution.

Measuring the time to initial display

The Android SDK automatically measures the TTID. The TTID can be optionally profiled using the Android Mobile Profiler.

Measuring the time to full display

The time to full display is manually defined using the GlobalRumMonitor.get().reportAppFullyDisplayed() API in the Android SDK based on the application’s specific definition of “fully drawn.”

Below is an example where time to full display is determined when home activity is fully loaded.

class HomeActivity : AppCompatActivity() {

    private val viewModel: HomeViewModel by viewModels()

    override fun onCreate(savedInstanceState: Bundle?) {
        setContentView(R.layout.activity_home)
        super.onCreate(savedInstanceState)

        viewModel.uiState.observe(this) { state ->
            if (state.isLoaded) {
                 GlobalRumMonitor.get().reportAppFullyDisplayed()
            }
        }
    }
}

If you use reportFullyDrawn to identify the moment of full display, you can use getFullyDrawnReporter to subscribe to reportFullyDrawn and call GlobalRumMonitor.get().reportAppFullyDisplayed().

RUM summary

The TTID and TTFD are presented in the RUM Summary under Mobile Performance. The standalone Mobile Performance Dashboard also contains distribution visuals for TTID and TTFD.

Vital events

The time to initial display and time to full display are presented as vital events in the RUM session. They are also found under the first view after the ApplicationLaunch view. TTID and TTFD are captured if the user launches the application in the session. Neither the TTID nor the TTFD appear if the user does not launch the application during the session.

The TTID and TTFD can be queried in the RUM Sessions Explorer using the following attributes on the vital event type:

• @vital.type:app_launch
• @vital.name:time_to_initial_display or @vital.name:time_to_full_display

Each TTID and TTFD side panel contains a distribution visualization, an indication of whether the launch was cold or warm, and an event waterfall.

Metrics

The time to initial display and time to full display are calculated as metrics:

• rum.measure.app.startup_to_initial_display, which represents the time to initial display
• rum.measure.app.startup_to_full_display, which represents the time to full display

These metrics contain the @vital.startup_type attribute to specify the launch type for accurate monitoring.

Further reading

추가 유용한 문서, 링크 및 기사:

Source code for dd-sdk-androidSOURCE CODE Explore Datadog RUMDOCUMENTATION