Event batching for the Java SDK

11 min

The Optimizely Feature Experimentation Java SDK batches decision and conversion events into a single payload before sending it to Optimizely. This is achieved through a new SDK component called the event processor.

Event batching has the advantage of reducing the number of outbound requests to Optimizely depending on how you define, configure, and use the event processor. It means less network traffic for the same number of Impression and conversion events tracked.

In the Java SDK, BatchEventProcessor provides implementation of the EventProcessor interface and batches events. You can control batching based on two parameters:

Batch size – Defines the number of events that are batched together before sending to Optimizely.

Flush interval – Defines the amount of time after which any batched events should be sent to Optimizely.

An event consisting of the batched payload is sent when the batch size reaches the specified limit or flush interval reaches the specified time limit. BatchEventProcessor options are described in more detail in the following section.

Note Event batching works with both out-of-the-box and custom event dispatchers.

The event batching process does not remove any personally identifiable information (PII) from events. You must ensure that you are not sending any unnecessary PII to Optimizely.

Basic example

If you want to use the default configuration of BatchEventProcessor (with the default AsyncEventHandler and HttpProjectConfigManager configurations), the SDK can be instantiated with the following OptimizelyFactory method. By default, batch size is 10 and flush interval is 30 seconds.

import com.optimizely.ab.Optimizely;
import com.optimizely.ab.OptimizelyFactory;

public class App {
    public static void main(String[] args) {
        String sdkKey = args[0];
        Optimizely optimizely = OptimizelyFactory.newDefaultInstance(sdkKey);
    }
}

Advanced Example

You can also customize BatchEventProcessor,AsyncEventHandler, and HttpProjectConfigManager using the builder options as shown in The following example:

import com.optimizely.ab.Optimizely;
import com.optimizely.ab.config.HttpProjectConfigManager;
import com.optimizely.ab.event.AsyncEventHandler;
import com.optimizely.ab.event.BatchEventProcessor;
import java.util.concurrent.TimeUnit;

public class App {
    public static void main(String[] args) {
        String sdkKey = args[0];
        
       ProjectConfigManager projectConfigManager = HttpProjectConfigManager.builder()
            .withSdkKey(sdkKey)
            .build();

        EventHandler eventHandler = AsyncEventHandler.builder()
            .withQueueCapacity(20000)
            .withNumWorkers(1)
            .build();

	// Here we are using the builder options to set batch size
        // to 50 events and flush interval to a minute.
        BatchEventProcessor batchProcessor = BatchEventProcessor.builder()
            .withBatchSize(50)
            .withEventHandler(eventHandler)
            .withFlushInterval(TimeUnit.MINUTES.toMillis(1))
            .build();

        Optimizely optimizely = Optimizely.builder()
            .withConfigManager(projectConfigManager)
            .withEventProcessor(batchProcessor)
            .build();
    }
}

Warning The maximum payload size is 3.5 MB. Optimizely rejects requests with a 400 response code, Bad Request Error, if the batch payload exceeds this limit.

The size limitation is because of the Optimizely Events API, which Feature Experimentation uses to send data to Optimizely.

The most common cause of a large payload size is a high batch size. If your payloads exceed the size limit, try configuring a smaller batch size.

BatchEventProcessor

BatchEventProcessor is an implementation of EventProcessor where events are batched. The class maintains a single consumer thread that pulls events off of the BlockingCollection and buffers them for either a configured batch size or a maximum duration before the resulting LogEvent is sent to the EventDispatcher and NotificationCenter.

The following properties can be used to customize the BatchEventProcessor configuration using the Builder class.

Builder methods

Use the following methods to customize the BatchEventProcessor.

Property	Default value	Description
withEventHandler	null	Required field. Specifies event handler to use to dispatch event payload to Optimizely. By default, if you create an instance, it passes in `AsyncEventHandler` to `BatchEventProcessor`.
withBatchSize	10	The maximum number of events to batch before dispatching. Once this number is reached, all queued events are flushed and sent to Optimizely.
withFlushInterval	30000 (30 Seconds)	Maximum time after which to batch and send event payload to Optimizely. In milliseconds.
withEventQueue	1000	BlockingCollection that queues individual events to be batched and dispatched by the executor.
withNotificationCenter	null	Notification center instance to be used to trigger any notifications to notify when event batches are flushed.
withExecutor	Single Thread Executor	Executor service that takes care of triggering event batching and subsequent dispatching to Optimizely.
withTimeout	5000	Maximum time to wait for event processor’s close to be executed. In milliseconds.

Advanced configuration

You can set the following properties to override the default configuration for BatchEventProcessor.

Property	Default value	Description
event.processor.batch.size	10	Maximum size of batch of events to send to Optimizely.
event.processor.batch.interval	30000	Maximum time after which to batch and send event payload to Optimizely. In milliseconds.
event.processor.close.timeout	5000	Maximum time to wait for event processor’s close to be executed. In milliseconds.

For information, see Initialize the Java SDK.

Side effects

The table lists other Optimizely functionality that may be triggered by using this class.

Functionality	Description
`LogEvent`	Whenever the event processor produces a batch of events, a `LogEvent` object is created using the `EventFactory`. It contains batch of conversion and decision events. This object is dispatched using the provided event dispatcher and also it is sent to the notification subscribers
Notification Listeners	Flush invokes the LOGEVENT [notification listener](https://docs.developers.optimizely.com/feature-experimentation/docs/set-up-notification-listener-java) if this listener is subscribed to.

Register a LogEvent listener

To register a LogEvent notification listener:

// There are two ways to register logEvent listener
// 1) Using NotificationCenter.
optimizely.notificationCenter.addNotificationHandler(LogEvent.class, logEventHandler -> {});

// 2) Using addLogEventNotificationHandler method of optimizely.
optimizely.addLogEventNotificationHandler(logEventHandler -> {
});

LogEvent

LogEvent object gets created using EventFactory. It represents the batch of decision and conversion events Optimizely sends to the Optimizely backend.

Object	Type	Description
requestMethod Required (non null)	RequestMethod (Enum)	The HTTP verb to use when dispatching the log event. It can be `GET` or `POST`.
endpointUrl\ Required (non null)	String	URL to dispatch log event to.
requestParams\ Required (non null)	Map\	Parameters to be set in the log event.
[eventBatch](https://github.com/optimizely/java-sdk/blob/13c27ffc812928ef5cf05159d37ad6bab421fc5a/core-api/src/main/java/com/optimizely/ab/event/LogEvent.java#L38)\ Required	EventBatch	It contains all the information regarding every event which is batched. including list of visitors which contains `UserEvent`.

Close Optimizely on application exit

If you enable event batching, you must call the Close method (optimizely.close()) before exiting. This ensures that queued events are flushed when possible to avoid data loss.

Warning Because the Optimizely client maintains a buffer of queued events, you must call close() on the Optimizely instance before shutting down your application or whenever dereferencing the instance.

Method	Description
close()	Stops all executor threads and flushes the event queue. This method will also stop any scheduledExecutorService that is running for the data-file manager.