Current Load - Warning / Critical

[experimenter]

Any suggestions on why my NEMS server is generating warnings and criticals for current load?  This is a Raspberry Pi 3 Model B, with only NEMS running on it.  Since I see lots of people running R-Pi3s and even R-Pi Zeros, it seems as if I should not be having load issues.

So far, I detected that it would spike after a reboot, as it looked like it was trying to query all the hosts at once at 15 minute intervals.  Since there are only six [6] hosts being monitored, I went in and manually changed the periods so that they are staggered [host1 = 12 min, host2 = 13 min, etc].

The 1, 5, 15 min parameters are set as follows:
Warning:  5.0, 4.0, 3.0
Critical:    10.0, 6.0, 4.0

I haven't saved all the emails it generated, but samples are:
WARNING - load average: 6.07, 5.50, 3.27
WARNING - load average: 3.79, 3.87, 3.64
WARNING - load average: 4.46, 4.26, 3.45

After some time it will settle down, and the post-revovery numbers look like this:
OK - load average: 1.32, 3.09, 2.88
OK - load average: 2.10, 2.31, 2.74

Obviously, I COULD raise the thresholds for triggering the warnings, but if I can determine that there is some way to lower the load, that would be preferable.

Thoughts?

[experimenter]

anybody???

[Robbie Ferguson]

Hi experimenter,
Can you provide a support.nems file so I can point you in the right direction?

It's not abnormal for a Raspberry Pi to have a pretty high load when working hard... it is what it is... but hitting a load average of 6 isn't something to be concerned about... just may want to adjust your alert thresholds for the NEMS server itself.

If your i7 was hitting a load average of 6, you'd wanna know. But a Pi hitting 6 is not overly abnormal (as long as it's happening in spikes, not 24/7).

Lemme know.

[experimenter]

OK, I'll get you the file.  Can you remind me where to locate it?

Thanks in advance,

[Robbie Ferguson]

You don't locate it, you create it.

https://docs.nemslinux.com/commands/nems-support

Cheers,
Robbie

[experimenter]

Thanks.  It's on its way to you.

For completeness, and as a reference to others who might read this thread: I did raise the thresholds and still got alerts.  After some correspondence with Robbie about relative safe levels for a Raspberry Pi to operate, I turned off monitoring the load.  If we can determine why this particular Pi is running so hard, I will gladly turn it back on.

FWIW, I have not yet turned on load monitoring for any other Pi servers running here, though that would be useful.

[Robbie Ferguson]

Yeah, the Pi is great, but not very powerful. It doesn't take much to bring it to a full load. Turning off the monitors is okay, though you may just need to find a more reasonable threshold that suits such a low-powered chip.
NEMS 1.5 will have new sample checks with much higher thresholds.
Unfortunately since you have disabled the load monitor, I cannot see the issue in your support.nems file (as it doesn't exists).
I do however see your weekly load average is just 1.5, so that's perfectly fine.
Perhaps sharing the notification output would help since it shows the load your system was hitting.
Cheers,
Robbie