Heron processing topologies can be written using an API called the Heron Streamlet API. The Heron Streamlet API is currently available for the following languages:

Although this document covers the new Heron Streamlet API, topologies created using the original topology API can still be used with Heron (which means that all of your older topologies will still run).

For a more in-depth conceptual guide to the new API, see The Heron Streamlet API. A high-level overview can also be found in the section immediately below.

The Heron Streamlet API vs. the topology API

When Heron was first released, all Heron topologies needed to be written using an API based on the Storm topology API. Although this API is quite powerful (and can still be used), the Heron Streamlet API enables you to create topologies without needing to implement spouts and bolts directly or to connect spouts and bolts together.

Here are some crucial differences between the two APIs:

Domain Original topology API Heron Streamlet API
Programming style Procedural, processing component based Functional
Abstraction level Low level. Developers must think in terms of "physical" spout and bolt implementation logic. High level. Developers can write processing logic in an idiomatic fashion in the language of their choice, without needing to write and connect spouts and bolts.
Processing model Spout and bolt logic must be created explicitly, and connecting spouts and bolts is the responsibility of the developer Spouts and bolts are created for you automatically on the basis of the processing graph that you build

The two APIs also have a few things in common:

  • Topologies' logical and physical plans are automatically created by Heron
  • Topologies are managed in the same way using the heron CLI tool

Getting started

In order to use the Heron Streamlet API for Java, you'll need to install the heron-api library.

Maven setup

In order to use the heron-api library, add this to the dependencies block of your pom.xml configuration file:


Compiling a JAR with dependencies

In order to run a Java topology created using the Heron Streamlet API in a Heron cluster, you'll need to package your topology as a "fat" JAR with dependencies included. You can use the Maven Assembly Plugin to generate JARs with dependencies. To install the plugin and add a Maven goal for a single JAR, add this to the plugins block in your pom.xml:


Once your pom.xml is properly set up, you can compile the JAR with dependencies using this command:

$ mvn assembly:assembly

By default, this will add a JAR in your project's target folder with the name PROJECT-NAME-VERSION-jar-with-dependencies.jar. Here's an example topology submission command using a compiled JAR:

$ mvn assembly:assembly
$ heron submit local \
  target/my-project-1.2.3-jar-with-dependencies.jar \
  com.example.Main \
  MyTopology arg1 arg2

Java Streamlet API starter project

If you'd like to up and running quickly with the Heron Streamlet API for Java, you can clone this repository, which includes an example topology built using the Streamlet API as well as the necessary Maven configuration. To build a JAR with dependencies of this example topology:

$ git clone https://github.com/streamlio/heron-java-streamlet-api-example
$ cd heron-java-streamlet-api-example
$ mvn assembly:assembly
$ ls target/*.jar

If you're running a local Heron cluster, you can submit the built example topology like this:

$ heron submit local target/heron-java-streamlet-api-example-latest-jar-with-dependencies.jar \
  io.streaml.heron.streamlet.WordCountStreamletTopology \

Selecting delivery semantics

Heron enables you to apply one of three delivery semantics to any Heron topology. For the example topology above, you can select the delivery semantics when you submit the topology with the topology's second argument. This command, for example, would apply effectively-once to the example topology:

$ heron submit local target/heron-java-streamlet-api-example-latest-jar-with-dependencies.jar \
  io.streaml.heron.streamlet.WordCountStreamletTopology \
  WordCountStreamletTopology \

The other options are at-most-once and at-least-once. If you don't explicitly select the delivery semantics, at-least-once semantics will be applied.

Streamlet API topology configuration

Every Streamlet API topology needs to be configured using a Config object. Here's an example default configuration:

import com.twitter.heron.streamlet.Config;
import com.twitter.heron.streamlet.Runner;

Config topologyConfig = Config.defaultConfig();

// Apply topology configuration using the topologyConfig object
Runner topologyRunner = new Runner();
topologyRunner.run("name-for-topology", conf, topologyBuilder);

The table below shows the configurable parameters for Heron topologies:

Parameter Default
Delivery semantics At most once
Serializer Kryo
Number of total container topologies 2
Per-container CPU 1.0
Per-container RAM 100 MB

Here's an example non-default configuration:

Config topologyConfig = Config.newBuilder()
        .setUserConfig("some-key", "some-value")

Delivery semantics

You can apply delivery semantics to a Streamlet API topology like this:


The other available options in the DeliverySemantics enum are ATMOST_ONCE and ATLEAST_ONCE.


In the Heron Streamlet API for Java, processing graphs consist of streamlets. One or more supplier streamlets inject data into your graph to be processed by downstream operators.


Operation Description Example
map Create a new streamlet by applying the supplied mapping function to each element in the original streamlet Add 1 to each element in a streamlet of integers
flatMap Like a map operation but with the important difference that each element of the streamlet is flattened Flatten a sentence into individual words
filter Create a new streamlet containing only the elements that satisfy the supplied filtering function Remove all inappropriate words from a streamlet of strings
union Unifies two streamlets into one, without modifying the elements of the two streamlets Unite two different Streamlet<String>s into a single streamlet
clone Creates any number of identical copies of a streamlet Create three separate streamlets from the same source
transform Transform a streamlet using whichever logic you'd like (useful for transformations that don't neatly map onto the available operations)
join Create a new streamlet by combining two separate key-value streamlets into one on the basis of each element's key Combine key-value pairs listing current scores (e.g. ("h4x0r", 127)) for each user into a single per-user stream
reduceByKeyAndWindow Produces a streamlet out of two separate key-value streamlets on a key, within a time window, and in accordance with a reduce function that you apply to all the accumulated values Count the number of times a value has been encountered within a specified time window
repartition Create a new streamlet by applying a new parallelism level to the original streamlet Increase the parallelism of a streamlet from 5 to 10
toSink Sink operations terminate the processing graph by storing elements in a database, logging elements to stdout, etc. Store processing graph results in an AWS Redshift table
log Logs the final results of a processing graph to stdout. This must be the last step in the graph.
consume Consume operations are like sink operations except they don't require implementing a full sink interface (consume operations are thus suited for simple operations like logging) Log processing graph results using a custom formatting function

Map operations

Map operations create a new streamlet by applying the supplied mapping function to each element in the original streamlet. Here's an example:

builder.newSource(() -> 1)
    .map(i -> i + 12);

In this example, a supplier streamlet emits an indefinite series of 1s. The map operation then adds 12 to each incoming element, producing a streamlet of 13s.

FlatMap operations

FlatMap operations are like map operations but with the important difference that each element of the streamlet is "flattened" into a collection type. In this example, a supplier streamlet emits the same sentence over and over again; the flatMap operation transforms each sentence into a Java List of individual words:

builder.newSource(() -> "I have nothing to declare but my genius")
    .flatMap((sentence) -> Arrays.asList(sentence.split("\\s+")));

The effect of this operation is to transform the Streamlet<String> into a Streamlet<List<String>>.

One of the core differences between map and flatMap operations is that flatMap operations typically transform non-collection types into collection types.

Filter operations

Filter operations retain elements in a streamlet, while potentially excluding some or all elements, on the basis of a provided filtering function. Here's an example:

builder.newSource(() -> ThreadLocalRandom.current().nextInt(1, 11))
        .filter((i) -> i < 7);

In this example, a source streamlet consisting of random integers between 1 and 10 is modified by a filter operation that removes all streamlet elements that are greater than 7.

Union operations

Union operations combine two streamlets of the same type into a single streamlet without modifying the elements. Here's an example:

Streamlet<String> oohs = builder.newSource(() -> "ooh");
Streamlet<String> aahs = builder.newSource(() -> "aah");

Streamlet<String> combined = oohs

Here, one streamlet is an endless series of "ooh"s while the other is an endless series of "aah"s. The union operation combines them into a single streamlet of alternating "ooh"s and "aah"s.

Clone operations

Clone operations enable you to create any number of "copies" of a streamlet. Each of the "copy" streamlets contains all the elements of the original and can be manipulated just like the original streamlet. Here's an example:

import java.util.List;
import java.util.concurrent.ThreadLocalRandom;

Streamlet<Integer> integers = builder.newSource(() -> ThreadLocalRandom.current().nextInt(100));

List<Streamlet<Integer>> copies = integers.clone(5);
Streamlet<Integer> ints1 = copies.get(0);
Streamlet<Integer> ints2 = copies.get(1);
Streamlet<Integer> ints3 = copies.get(2);
// and so on...

In this example, a streamlet of random integers between 1 and 100 is split into 5 identical streamlets.

Transform operations

Transform operations are highly flexible operations that are most useful for:

  • operations involving state in stateful topologies
  • operations that don't neatly fit into the other categories or into a lambda-based logic

Transform operations require you to implement three different methods:

  • A setup method that enables you to pass a context object to the operation and to specify what happens prior to the transform step
  • A transform operation that performs the desired transformation
  • A cleanup method that allows you to specify what happens after the transform step

The context object available to a transform operation provides access to:

  • the current state of the topology
  • the topology's configuration
  • the name of the stream
  • the stream partition
  • the current task ID

Here's a Java example of a transform operation in a topology where a stateful record is kept of the number of items processed:

import com.twitter.heron.streamlet.Context;
import com.twitter.heron.streamlet.SerializableTransformer;

import java.util.function.Consumer;

public class CountNumberOfItems implements SerializableTransformer<String, String> {
    private int numberOfItems;

    public void setup(Context context) {
        numberOfItems = (int) context.getState("number-of-items");
        context.getState().put("number-of-items", numberOfItems + 1);

    public void transform(String in, Consumer<String> consumer) {
        String transformedString = // Apply some operation to the incoming value

    public void cleanup() {
                String.format("Successfully processed new state: %d", numberOfItems));

This operation does a few things:

  • In the setup method, the Context object is used to access the current state (which has the semantics of a Java Map). The current number of items processed is incremented by one and then saved as the new state.
  • In the transform method, the incoming string is transformed in some way and then "accepted" as the new value.
  • In the cleanup step, the current count of items processed is logged.

Here's that operation within the context of a streamlet processing graph:

builder.newSource(() -> "Some string over and over");
        .transform(new CountNumberOfItems())

Join operations

For a more in-depth conceptual discussion of joins, see the Heron Streamlet API doc.

Join operations unify two streamlets on a key (join operations thus require KV streamlets). Each KeyValue object in a streamlet has, by definition, a key. When a join operation is added to a processing graph,

import com.twitter.heron.streamlet.WindowConfig;

Builder builder = Builder.CreateBuilder();

KVStreamlet<String, String> streamlet1 =
        builder.newKVSource(() -> new KeyValue<>("heron-api", "topology-api"));

builder.newSource(() -> new KeyValue<>("heron-api", "streamlet-api"))
    .join(streamlet1, WindowConfig.TumblingCountWindow(10), KeyValue::create);

In this case, the resulting streamlet would consist of an indefinite stream with two KeyValue objects with the key heron-api but different values (topology-api and streamlet-api).

The effect of a join operation is to create a new streamlet for each key.

Reduce by key and window operations

You can apply reduce operations to streamlets by specifying:

  • a key extractor that determines what counts as the key for the streamlet
  • a value extractor that determines which final value is chosen for each element of the streamlet
  • a time window across which the operation will take place
  • a reduce function that produces a single value for each key in the streamlet

Reduce by key and window operations produce a new streamlet of key-value window objects (which include a key-value pair including the extracted key and calculated value, as well as information about the window in which the operation took place). Here's an example:

import java.util.Arrays;

import com.twitter.heron.streamlet.WindowConfig;

Builder builder = Builder.newBuilder()
    .newSource(() -> "Mary had a little lamb")
    // Convert each sentence into individual words
    .flatMap(sentence -> Arrays.asList(sentence.toLowerCase().split("\\s+")))
        // Key extractor (in this case, each word acts as the key)
        word -> word,
        // Value extractor (each word appears only once, hence the value is always 1)
        word -> 1,
        // Window configuration
        // Reduce operation (a running sum)
        (x, y) -> x + y
    // The result is logged

Repartition operations

When you assign a number of partitions to a processing step, each step that comes after it inherits that number of partitions. Thus, if you assign 5 partitions to a map operation, then any mapToKV, flatMap, filter, etc. operations that come after it will also be assigned 5 partitions. But you can also change the number of partitions for a processing step (as well as the number of partitions for downstream operations) using repartition. Here's an example:

import java.util.concurrent.ThreadLocalRandom;

Builder builder = Builder.CreateBuilder();

builder.newSource(() -> ThreadLocalRandom.current().nextInt(1, 11))
        .map(i -> i + 1)
        .filter(i -> i > 7 && i < 2)

In this example, the supplier streamlet emits random integers between one and ten. That operation is assigned 5 partitions. After the map operation, the repartition function is used to assign 2 partitions to all downstream operations.

Sink operations

In processing graphs like the ones you build using the Heron Streamlet API, sinks are essentially the terminal points in your graph, where your processing logic comes to an end. A processing graph can end with writing to a database, publishing to a topic in a pub-sub messaging system, and so on. With the Streamlet API, you can implement your own custom sinks. Here's an example:

import com.twitter.heron.streamlet.Context;
import com.twitter.heron.streamlet.Sink;

public class FormattedLogSink implements Sink<T> {
    private String streamletName;

    public void setup(Context context) {
        streamletName = context.getStreamletName();

    public void put(T element) {
        String message = String.format("Streamlet %s has produced an element with a value of: '%s'",

    public void cleanup() {}

In this example, the sink fetches the name of the enclosing streamlet from the context passed in the setup method. The put method specifies how the sink handles each element that is received (in this case, a formatted message is logged to stdout). The cleanup method enables you to specify what happens after the element has been processed by the sink.

Here is the FormattedLogSink at work in an example processing graph:

Builder builder = Builder.newBuilder();

builder.newSource(() -> "Here is a string to be passed to the sink")
        .toSink(new FormattedLogSink());

Log operations rely on a log sink that is provided out of the box. You'll need to implement other sinks yourself.

Log operations

Log operations are special cases of consume operations that log streamlet elements to stdout.

Streamlet elements will be using their toString representations and at the INFO level.

Consume operations

Consume operations are like sink operations except they don't require implementing a full sink interface. Consume operations are thus suited for simple operations like formatted logging. Here's an example:

Builder builder = Builder.newBuilder()
        .newSource(() -> generateRandomInteger())
        .filter(i -> i % 2 == 0)
        .consume(i -> {
            String message = String.format("Even number found: %d", i);