Hi there,
I would like to expand on this knowledge base article which goes into detail on how the results from Linux System Load Sensors like "SNMP Linux Load Average are to be interpreted: https://kb.paessler.com/en/topic/43993-what-is-the-load-on-linux-systems
According to this article the load value can surpass 1 and even higher numbers, if there are mutliple processing cores. But how does this sensor behave in systems with multiple Sockets, i.e. multiple CPU´s with each having multiple cores? Specifically I would like to get load data from a Citrix XenServer Host with 2 CPU Sockets, each having 24 cores. If the load value behaves the same, in my mind i´d have to add up all the available cores over all sockets, and as long as the load stays under that number or under 70% of that respectively the system is not overloaded. Or is it possible that the load value in this scenario now represents each socket, and in a two socket system it can now only max out at 2? I´m asking because the load from this specific host very rarely surpasses 1, that´s why I suspect the value is representing the socket count here rather than the actual core count.
Thanks for sheding some light on this.
Add comment