Skip to content

Latest commit

 

History

History
229 lines (168 loc) · 8.04 KB

android.md

File metadata and controls

229 lines (168 loc) · 8.04 KB
Raw
title description keywords author ms.author ms.date ms.topic ms.assetid ms.service ms.custom ms.tgt_pltfrm dev_langs
App Center Analytics for Android
App Center Analytics for Android
analytics
lucen-ms
lucen
11/26/2021
article
5392ac23-465d-464d-a533-262a94cf15c3
vs-appcenter-sdk
sdk
android
java
kotlin

App Center Analytics (Android)

[!INCLUDE Retirement announcement for App Center]

[!div class="op_single_selector"]

App Center Analytics helps you understand user behavior and customer engagement to improve your app. The SDK automatically captures session count and device properties like model, OS version, etc. You can define your own custom events to measure things that matter to you. All the information captured is available in the App Center portal for you to analyze the data.

Follow the SDK Getting started section if you haven't set up the SDK in your application yet.

Session and device information

Once you add App Center Analytics to your app and start the SDK, it will automatically track sessions and device properties like OS Version, model, etc. without writing any additional code.

Country Code

The SDK automatically reports a user's country code if the device has a mobile data modem and a SIM card installed. WiFi-only devices won't report a country code by default. To set the country code of those users, you must retrieve your user's location yourself and use the setCountryCode: method in the SDK:

AppCenter.setCountryCode("en");
AppCenter.setCountryCode("en")

Note

For country code to be displayed on Analytics sessions, AppCenter.setCountryCode must be called prior to calling AppCenter.start.

Custom events

You can track your own custom events with up to 20 properties to understand the interaction between your users and the app.

Once you've started the SDK, use the trackEvent() method to track your events with properties. You can send up to 200 distinct event names. Also, there's maximum character limits:

  • 256 characters per event name.
  • 125 characters per event property name & event property value.
Map<String, String> properties = new HashMap<>();
properties.put("Category", "Music");
properties.put("FileName", "favorite.avi");

Analytics.trackEvent("Video clicked", properties);
val properties = hashMapOf("Category" to "Music", "FileName" to "favorite.avi")
Analytics.trackEvent("Video clicked", properties)

Properties for events are entirely optional – if you just want to track an event, use this sample instead:

Analytics.trackEvent("Video clicked");
Analytics.trackEvent("Video clicked")

Event priority and persistence

You can track business critical events that have higher importance than other events.

  • Developers can set priority of events as Normal (Flags.NORMAL in the API) or Critical (Flags.CRITICAL in the API).
  • Events with priority set as Critical will be retrieved from storage first and sent before Normal events.
  • When the local storage is full, and new events need to be stored, the oldest events with the lowest priority are deleted first.
  • If the storage is full of logs with Critical priority, then tracking an event with Normal priority will fail as the SDK can't make room in that case.
  • If you also use the Crashes service, crash logs are set as Critical and share the same storage as events.
  • The transmission interval is only applied to Normal events, Critical events will be sent after 3 seconds.

You can use the following API to track an event as Critical:

Map<String, String> properties = new HashMap<>();
properties.put("Category", "Music");
properties.put("FileName", "favorite.avi");

Analytics.trackEvent("eventName", properties, Flags.CRITICAL);

// If you're using name only, you can pass null as properties.
val properties = hashMapOf("Category" to "Music", "FileName" to "favorite.avi")
Analytics.trackEvent("Video clicked", properties, Flags.CRITICAL)

// If you're using name only, you can pass null as properties.

Pause and resume sending logs

Pausing the event transmission can be useful in scenarios when the app needs to control the network bandwidth for more business critical needs. You can pause sending logs to the App Center backend. When paused, events can still be tracked and saved, but they aren't sent right away. Any events your app tracks while paused will only be sent once you call resume.

Analytics.pause();
Analytics.resume();
Analytics.pause()
Analytics.resume()

Enable or disable App Center Analytics at runtime

You can enable and disable App Center Analytics at runtime. If you disable it, the SDK won't collect any more analytics information for the app.

Analytics.setEnabled(false);
Analytics.setEnabled(false)

To enable App Center Analytics again, use the same API but pass true as a parameter.

Analytics.setEnabled(true);
Analytics.setEnabled(true)

The state is persisted in the device's storage across application launches.

[!INCLUDE android see async]

Note

This method must only be used after Analytics has been started.

Check if App Center Analytics is enabled

You can also check if App Center Analytics is enabled or not.

Analytics.isEnabled();
Analytics.isEnabled()

[!INCLUDE android see async]

Note

This method must only be used after Analytics has been started, it will always return false before start.

[!INCLUDE manual session tracker]

  • Call the following method before the SDK start:
Analytics.enableManualSessionTracker();
Analytics.enableManualSessionTracker()
  • Then you can use the startSession API after the AppCenter.start:
Analytics.startSession();
Analytics.startSession()

Local storage size

By default, the SDK stores all the event logs up to 10 MB. Developers can use an API to increase the storage size and the SDK will keep storing logs until the storage is full.

No internet access

When there's no network connectivity, the SDK saves up to 10 MB of logs in the local storage. Once the storage is full, the SDK starts discarding old logs to make room for the new logs. Once network connectivity returns, the SDK sends logs in the batch of 50 or after every 6 seconds (by default).

Note

The logs older than 25 days won't be accepted by the backend.

Batching event logs

The App Center SDK uploads logs in a batch of 50 and if the SDK doesn't have 50 logs to send, it will still send logs after 6 seconds (by default). There can be a maximum of three batches sent in parallel. The transmission interval can be changed:

// Change transmission interval to 10 seconds.
Analytics.setTransmissionInterval(10000);
// Change transmission interval to 10 seconds.
Analytics.setTransmissionInterval(10000)

The transmission interval value must be between 6 seconds and 86400 seconds (one day) and this method must be called before the service is started.

Retry and back-off logic

App Center SDK supports back-off retries on recoverable network errors. Below is the retry logic:

  • 3 tries maximum per request.
  • Each request has its own retry state machine.
  • All the transmission channels are disabled (until next app process) after one request exhausts all its retries.

Back-off logic

  • 50% randomization, first retry between 5 and 10 seconds, next try between 2.5 and 5 minutes, last try between 10 and 20 minutes.
  • If network switches from off to on (or from wi-fi to mobile), retry states are reset and requests are retried immediately.