Skip to main content

Pekko Runtime

PekkoRuntime leverages Pekko Persistence and Cluster Sharding to provide a distributed runtime for workflows.

PekkoRuntime operates natively with scala.concurrent.Future — your workflow context should declare type Effect[T] = Future[T] or lift the desired effect as described in effect types.

"org.business4s" %% "workflows4s-pekko" % "0.6.0"

Example​

Here is an example of using PekkoRuntime in a workflow:

import MyWorkflowCtx.*
given ActorSystem[?]                            = ???
val engine: WorkflowInstanceEngine[Future, Ctx] = ???
val workflow: WIO.Initial                       = ???

val runtime: PekkoRuntime[Ctx] = PekkoRuntime.create("my-workflow", workflow, InitialState(), engine)

runtime.initializeShard()

val instance: WorkflowInstance[Future, State] = runtime.createInstance_("my-workflow-id")

Caveats​

While PekkoRuntime offers robust distributed capabilities, it comes with some caveats:

  1. Configuration Complexity:
    • A significant amount of complexity is encapsulated within Pekko.
    • Proper configuration is essential to ensure smooth operation.
  2. Initial State Limitation:
    • Unlike other runtimes, PekkoRuntime does not support arbitrary initial states during workflow instance creation. Initial state has to be build uniformly across all instances based on EntityContext.
  3. Shard Initialization:
    • The initializeShard method must be invoked before creating workflow instances to ensure proper setup.