datadog_flutter 1.1.1

Community implementation of native bindings for Datadog's SDK. Not an official package.

Datadog Flutter

Community implementation of native bindings for Datadog's SDK. This is not an official package.

Setup

1. Generate a client token from Datadog through the Settings > API panel (under Client Tokens).
2. Initialize:

   await DatadogFlutter.initialize(
     clientToken: myDatadogClientToken,
     serviceName: 'my-app-name',
     environment: 'production',
   )
   

3. Associate RUM and log events (optional):

   await DatadogFlutter.setUserInfo(id: <YOUR_USER_ID>);
   

⚠️ Your Podfile must have use_frameworks! (Flutter includes this by default) and your minimum iOS target must be >= 11. This is a requirement from the Datadog SDK.

Logging

In its default implementation, log data will only be transmitted to Datadog through Logger records. print statements are not guaranteed to be captured.

ddLogger = DatadogLogger(loggerName: 'orders');
// optionally set a value for HOST
// ddLogger.addAttribute('hostname', <DEVICE IDENTIFIER>);

ddLogger.addTag('restaurant_type', 'pizza');
ddLogger.removeTag('restaurant_type');

// add attribute to every log
ddLogger.addAttribute('toppings', 'extra_cheese');

// add atttributes to some logs
ddLogger.log('time to cook pizza', Level.FINE, attributes: {
  'durationInMilliseconds': timer.elapsedMilliseconds,
});

Real User Monitoring

RUM adds support for error, event, and screen tracking. The integration requires additional configuration for each service.

1. Supply an application ID to initialize:

   await DatadogFlutter.initialize(
     clientToken: myDatadogClientToken,
     serviceName: 'my-app-name',
     environment: 'production',
     iosRumApplicationId: myiOSRumApplicationId,
     androidRumApplicationId: myAndroidRumApplicationId,
   )
   

2. Acknowledge TrackingConsent at initialization or later within your application. Events will not be logged until trackingConsent is .granted. This value can be updated via DatadogFlutter.updateTrackingConsent.
3. Automatically track screens:

   MaterialApp(
     // ...your material config...
     home: HomeScreen(),
     navigatorObservers: [
       DatadogObserver(),
     ],
   );
   

4. Automatically report errors:

   void main() async {
     // Capture Flutter errors automatically:
     FlutterError.onError = DatadogRum.instance.addFlutterError;
   
     // Catch errors without crashing the app:
     runZonedGuarded(() {
       runApp(MyApp());
     }, (error, stackTrace) {
       DatadogRum.instance.addError(error, stackTrace);
     });
   }
   

5. Manually track additional events:

   GestureDetector(onTap: () {
     DatadogRum.instance.addUserAction('EventTapped');
   })
   

6. Manually track additional errors:

   try {
     throw StateError();
   } catch (e, st) {
     DatadogRum.instance.addError(e, st);
   }
   

7. Manually track network requests or resources:

   await DatadogRum.startResourceLoading(
     aUniqueIdentifier,
     url: 'https://example.com',
     method: RUMResources.get,
   );
   await DatadogRum.stopResourceLoading(
     aUniqueIdentifier,
     statusCode: 500,
     errorMessage: 'Internal Server Error' ,
   )
   

Tracing

Associate your HTTP requests with your backend service. Be sure to setup (usually immediately after DatadogFlutter.initialize):

await DatadogTracing.initialize();

For one-off requests, instantiate a fresh client:

final httpClient = DatadogTracingHttpClient();
// make requests
final response = await httpClient.get('http://example.com');

For frameworks that use an internal client like Brick or Dio, compose the client:

RestProvider(
  client: DatadogTracingHttpClient();
)
// or compose if another client is already being used:
RestProvider(
  client: DatadogTracingHttpClient(GZipHttpClient());
)

FAQ

How do I disable logging when I'm developing locally?

By default, the DatadogFlutter default constructor will send all logs from Logger to Datadog. To ignore, set bindOnRecord:

DatadogLogger(bindOnRecord: false)

And log conditionally later:

Logger.root.onRecord.listen((record) async {
  if (shouldSendToDatadog) {
    ddLogger.onRecordCallback(record)
  } else {
    print(record);
  }
});