Maybe a dumb question, but I posted this screenshot of my tests of Topbeat on twitter
The CPU usage on the process I picked the perc value expecing it to be 0-100, but the number on the graph goes much higher, to 180. How is this counted? is it per core, per socket or ?
Topbeat takes the data from the top command which can also show more then 100%. As far as I know, in case you have 4 cores and 3 of them are used 40%, the output will bet 120%.
I see, very informative but makes me wonder. When you have statistics like that coming in, then writing a monitoring rule on top of that becomes difficult because the values you put in the rule change depending on how many cores the endpoint has. Is this something to worry over?
In hindsight when looking at historical metrics like this its very good to see values for individual cores. For the future would it be possible to include usage per core instead of the total?
hostname.proc.procname.cpu0, hostname.proc.procname.cpu1 and so on.
That is actually a good point. I can't say if this is possible for Windows, but for Linux it should be doable as this information can also be fetched from top or other commands. If you open the top command and press 1, it shows the use per CPU.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.