Dealing with top-level Actor exceptions

Akka actors are great for reactively and asynchronously handling work loads. Typically, for larger actor systems, it’s best to use a Supervisor to handle unexpected failure. However, for top-level actors, it’s nice to be able to intercept and handle exceptions for alerting purposes. For this specific purpose, we created an actor trait called AlertingActor, which very simply defines a preRestart that gets called with the Throwable when an unhandled exception occurs in an actor.

trait AlertingActor extends Actor {
  def alert(reason: Throwable, message: Option[Any]) = {
    // hook into your pre-existing alerting / healthcheck system
    //  i.e., send emails, log, or call an error API
  }
 
  override def preRestart(reason: Throwable, message: Option[Any]) {
    alert(reason, message)
    postStop()
  }
}
 
// and in use:
class SomeActor extends AlertingActor {
  // ...
}