[akka-user] Confused by continuous stream of DistributedPubSub dead letters in Akka Cluster system

Hi, We have an Akka Cluster system with 4 nodes (with 2 different roles: 3 frontend nodes and 1 backend node). We are confused by a continuous stream of dead letters that we have been observing, in the backend node. They occur at a rate of about 20 to 30 per minute, per frontend. The sender of

[akka-user] Re: Terminated messages arriving less with congested thread pool?

Nevermind, monitorizing with Kamon we've seen that some mailboxes were growing too large. Thanks El dilluns, 4 abril de 2016 13:51:08 UTC+2, Artur Soler va escriure: > > Hi, > > Our application has a pool of workers processing events, which trigger the > start

[akka-user] Terminated messages arriving less with congested thread pool?

Hi, Our application has a pool of workers processing events, which trigger the start of some actors that automatically stop themselves after some checks, and some other actors which have watched them react on that. We're finding some unexpected behavior in that, when the workers pool is small