Open
Description
Hello,
Doing some tests I've found that case the logstash server is unreachable for some reason, the current implementation would simply stash everything within the MetricShipper actor, which in the worst case leads to an OutOfMemoryException.
As it is not expected that a monitoring solution impact the monitored system, I would expect stashing messages to be optional at least.
Architecturally there are different ways to model this option, but as I don't have the real need for it and I'm new to this library, maybe @darienmt would have a better/quicker to implement idea.
Meanwhile in my fork I just disabled that completely.
Cheers,
Oswaldo
Metadata
Metadata
Assignees
Labels
No labels