Hello,
Is there any document with the meaning of the statistics, like for example what does EndpointLostCount means or WorkUnitProcessingRate and the meaning of the"father" statistic like: CEP-Engine, JVM.Memory and so on....
Hello,
Is there any document with the meaning of the statistics, like for example what does EndpointLostCount means or WorkUnitProcessingRate and the meaning of the"father" statistic like: CEP-Engine, JVM.Memory and so on....
I don't get how is it possible for the vendor making a product not having documentation about their product. It's beyond me.
The description field in the stats browser should help a little.
The WorkGroup/WorkUnit stats for the Log Collection modules are indicators of the internal process state and normally would be used only for debugging but a few are useful for checking if a collection is not performing well. For example the "WorkUnits Queued Wait Time", is an indication of how long something is queued to be processed before execution occurs. "WorkUnits Queued" indicates the number of eventsources queued for execution.
Not sure about "EndpointLostCount". The "Health Checks" are indicators of how the service is doing, I agree that the description could be improved.
In terms of overall status and processing of each service there is a composite statistic for this which is used to drive the lights on the H&W monitoring page. These wrap up a lot of main issues with each service.
Hello Scott,
Thanks for the reply.
Looking in Syestem Stats Browser we can get a picture of the service, but only the picture not the meaning. For example: In can the that process.jvm.memory-health is related with the health of something in the server, probably the Java virutal machine, but what happens it this field is unhealthy? How, and wich way, does it affect the server? This are just some of the questions i have....
For example Scoot,
I want to monitor with a policie the ESA service in the category filesystem and has far i can see its not possible.
Hello Marinos,
Thats it. I thought that all the policies were made in the respective component we wanted to monitor. I can see now that the policies are made in "mysterious ways".
For anything related to the service itself, it will be in the respective component as you said. For anything hardware, sensor, OS etc it will be under Hosts. You can play a little with the dropdowns in Health and Wellness to see what is recorded and where, then you can create your policies.
In H&W for example, select the ESA hostname under Host and select Host under component to see all the "Host" related stats that are recorded for that ESA box.
Getting a detailed explanation of each of these stats is another story and we may never know:)
For my understanding the second field in the system stats browser is the one that gives us the "appliance" were we can find the statistic. I'll work with that .
Thanks
PS: DO you have your ESA rules sending email alerts by any chance?
Yeah it's confusing if you didn't know this already.
No, I don't have esa rules for this. You can configure e-mail alerts directly from the policy as well as syslog alerts which can be fed back to a Log decoder/VLC.
I wouldn't personally configure ESA alerts for this because it will add unnecessary complexity and add (a little) more load on ESA. However the choice is yours and I don't know your general setup and how you prefer to have your alerts.
Anyone?