Get monitoring, observability, online education,
and expert support from Lightbend.
Learn More

Debugging Streamlets

Debugging runLocal

By default Cloudflow runs on debug mode when runLocal. In the output we can see the listening port for each Streamlet.

[info] listening for debugging '[streamlet-name]' at 'localhost:[port]'

We can attach to this process by adding a remote configuration in the IDE of our choice. This Run/Debug configuration for the JVM must have the same port of the Streamlet we want to debug. See the following for an example configuration:


Debugging Remote

Debugging a remote application is possible just by applying the standard configuration to debugging a remote JVM. That is, adding a conf like this to JAVA_OPTS through our configuration model before we deploy or configure our Cloudflow app.


As we’ve seen in the Configuration model, adding these parameters to JAVA_OPTS can be done at the Streamlet level or runtime level. In this example, we are adding it to all Akka streamlets.

cloudflow.akka.runtime {
  kubernetes.pods.pod {
    containers.container {
      env = [
            { name = "JAVA_OPTS"
              value = "-agentlib:jdwp=transport=dt_socket,server=y,suspend=n,address=5005"

Once this is added, and the application is deployed or configured, we’ll need to do two things:

  1. Add a remote configuration in the IDE of our choice. This remote Run/Debug configuration for the remote JVM must have the same arguments, except for the address set in JAVA_OPTS. The address can be different because it belongs to the environment where the IDE is running, typically will be local. See the following for an example configuration:

  2. Port-forward a local address to the selected streamlet/pod in the Kubernetes cluster. Let’s say our application is called 'example-application' and we picked the above parameters for the IDE debug configuration. Then we will need to port-forward like the following:

    kubectl port-forward pod/[some-streamlet-pod-name] 4004:5005 -n example-application