Getting Started with Samza REST

This tutorial depends on hello-samza to start some example jobs on a local cluster, which you will then access via the JobsResource. After completing this tutorial, you will have built and deployed the Samza REST resource locally, changed the configuration for the JobsResource, and executed a couple of basic curl requests to verify the service works.

Lets get started.

Run Hello Samza Jobs Locally

Follow the hello-samza tutorial to setup a local grid and run the wikipedia jobs. Skip the shutdown step because you need the grid to still be running to query the REST service for jobs. You can optionally skip all the commands if you don’t want to verify the output of the jobs.

Take note of the path where you cloned hello-samza. You will need this to configure the installations path for the JobsResource.

Build the Samza REST Service package

The source code for Samza REST is in the samza-rest module of the Samza repository. It To build it, execute the following gradle task from the root of the project.

./gradlew samza-rest:clean releaseRestServiceTar

Deploy the Samza REST Service Locally

To deploy the service, you simply extract the tarball to the desired location. Here, we will deploy the tarball on the local host in


where SAMZA_ROOT is the path to the root of your Samza project.

Run the following commands:

cd samza-rest/build/distributions/
mkdir -p deploy/samza-rest
tar -xvf ./samza-rest-0.11.0-SNAPSHOT.tgz -C deploy/samza-rest

Configure the Installations Path

The JobsResource has a required config job.installations.path which specifies the path where the jobs are installed. Edit the configuration file:


Set the job.installations.path to:


where hello-samza-ROOT is the path to your hello-samza clone, noted above. This tells the JobsResource to crawl this location to find all the installed jobs.

Start the Samza REST Service

To deploy the service, run the script from the extracted directory.

cd deploy/samza-rest
./bin/  \
  --config-factory=org.apache.samza.config.factories.PropertiesConfigFactory \

You provide two parameters to the script. One is the config location, and the other, optional, parameter is a factory class that is used to read your configuration file. The SamzaRestService uses your ConfigFactory to get a Config object from the config path. The ConfigFactory is covered in more detail on the Job Runner page. The script will block until the SamzaRestService terminates.

Note: With the default settings, the JobsResource will expect a YARN cluster with a local Resource Manager accessible via the ApplicationCLI. Without YARN, the JobsResource will not respond to any requests. So it’s important to walk through hello-samza demo before the next step.

Curl the Default REST Service

Curl the JobsResource to get all installed jobs

curl localhost:9139/v1/jobs

Now curl the JobsResource to stop one of the jobs

curl -X PUT localhost:9139/v1/jobs/wikipedia-feed/1?status=stopped

Congratulations, you’ve successfully deployed the Samza REST Service and used the JobsResource to list jobs and stop a job!

See the JobsResource documentation for the rest of its API.

See the Resources documentation for more information about Resources and how you can add your own.