I'm doing an alert about "network connection refused by other systems" for my customer.
They told me some of their external systems are very unstable and they want to receive alert when these systems fail to response to java application.
I try to use Introscope Agent to issue this alert so I modified errors.pbd to add these settings.
TraceOneMethodWithParametersOfClass: java.net.Socket connect BlamePointTracer "Important Connection"
TraceOneMethodWithParametersOfClass: java.net.Socket connect ExceptionErrorReporter "Important Connection:Errors Per Interval"
Yes I try to answer this question by receiving exception from java.net.Socket. (connectException)
After reboot the agent, I found that I can receive exception and error count also increased but I can't figure out which external system failed by their ip address. They all showed within the metric group named "Important Connection".
Is there any way to group the metrics by their ip address and port?