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
I think a lot of people would appreciate a Flow that would seamlessly help dealing with strange throughput decreases/issues. Either in form of logging & profiling capabilities or throwing exceptions that would reveal the culprit.
Imagine having a stream of 10 components using 5 different remote services and the stream suddenly looses 50% of original throughput, there is almost no way of targeting this problem right now which leads to a whole day of troubleshooting because you don't know whether it is caused by the services, by akka version upgrade or by a new feature you added. It's madness unless you are an akka-stream guru :-)
Within a stream we can at best create combinators that measure throughput and allow reactions to the measured values, but in a back-pressured system this will not tell you where the problem lies—it could even be downstream. This concern will need to addressed by external holistic monitoring tools.
Hey,
I think a lot of people would appreciate a Flow that would seamlessly help dealing with strange throughput decreases/issues. Either in form of logging & profiling capabilities or throwing exceptions that would reveal the culprit.
Imagine having a stream of 10 components using 5 different remote services and the stream suddenly looses 50% of original throughput, there is almost no way of targeting this problem right now which leads to a whole day of troubleshooting because you don't know whether it is caused by the services, by akka version upgrade or by a new feature you added. It's madness unless you are an akka-stream guru :-)
Based on these 2 discussions :
stream-idling-on-suddenly-slowed-down-services
profiling-options
The text was updated successfully, but these errors were encountered: