You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
Similar to how we implemented some built-in telemetry for actor starts / stops in #6293, we've had a request on our Phobos issue tracker to do the same for some of the moving parts inside Akka.Remote and Akka.Cluster: petabridge/phobos-issues#79 as part of Phobos
I think we can probably achieve this - these metrics are already captured inside the remoting and clustering systems, but they're not exposed in any meaningful way that could easily be consumed for instrumentation purposes.
Describe the solution you'd like
I think we should create topics for each of the major heartbeat systems:
Akka.Remote - Transport
Akka.Remote - DeathWatch
Akka.Cluster - Watch
and one for Akka.Remote - transport metrics.
And make these subscribable via the EventStream locally, for that node's traffic only.
Describe alternatives you've considered
The alternatives in this case are basically "not doing it" - or trying to do something really janky in Phobos (i.e. custom failure detector registries.) Doing this natively in Akka.NET is the way to do it - and this won't have much of a performance impact since these messages would only be shared once every 5s or so.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
Similar to how we implemented some built-in telemetry for actor starts / stops in #6293, we've had a request on our Phobos issue tracker to do the same for some of the moving parts inside Akka.Remote and Akka.Cluster: petabridge/phobos-issues#79 as part of Phobos
I think we can probably achieve this - these metrics are already captured inside the remoting and clustering systems, but they're not exposed in any meaningful way that could easily be consumed for instrumentation purposes.
Describe the solution you'd like
I think we should create topics for each of the major heartbeat systems:
and one for Akka.Remote - transport metrics.
And make these subscribable via the
EventStream
locally, for that node's traffic only.Describe alternatives you've considered
The alternatives in this case are basically "not doing it" - or trying to do something really janky in Phobos (i.e. custom failure detector registries.) Doing this natively in Akka.NET is the way to do it - and this won't have much of a performance impact since these messages would only be shared once every 5s or so.
The text was updated successfully, but these errors were encountered: