Thank you and a note about system load average #96
tonymartino
started this conversation in
Ideas
Replies: 1 comment 7 replies
-
Good notes! I’ll look into updating the code to better reflect reality. |
Beta Was this translation helpful? Give feedback.
7 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Thanks for putting this together -- I'm a new user and am already getting some utility out of it. I was wondering why the system load average values were being displayed as a percentage in my entity card and found that you've set the unit_of_measurement attribute to %. Please note that the load average is not a percentage value but rather a calculated value indicating the number of processes in the running or runnable states that are waiting for CPU time. The Wikipedia article has a very good explanation of this. Would you consider removing the unit_of_measure attribute from the load average entities?
Another important aspect of system load average is that it is not, in fact, CPU load average. Processes can be waiting for CPU time even when the system's CPUs are under no load at all. For example, something I've seen on many systems over the years (and a scenario the Wikipedia article discusses briefly) is NFS failures. A system can reach a load average in the hundreds when critical IO is blocked or slow. This leaves CPUs idling while processes in the running state can not be processed by the CPU because the system is waiting for date. Indeed, the screenshot here shows a system with ~10% CPU utilization and a load average reaching 108! So, I'm also hoping you'd consider changing the entity names from %systemname% CPU Load Average* to %systemname% System Load Average* or, more simply, %systemname Load Average*. Thanks for your consideration and thanks again for the integration!
Beta Was this translation helpful? Give feedback.
All reactions