Struggling with CPU

[craiga_uk]

Hi all,

Hopefully some experts here can help me. We have a number (approx 20) Untangle devices out on sites, supporting between 25 - 200 devices.

In recent times (the last year or so) we've been exclusively using Untangle appliances, which work well.
The older systems we have are all built on HP Gen 8 Microservers, most of them running Celeron G1610T processors:

Server: custom
CPU Count: 2
CPU Type: Intel(R) Celeron(R) CPU G1610T @ 2.30GHz
Architecture: amd64
Memory: 4 GB
Disk: 933.75 GB

We're hitting random issues on some sites with high CPU alerts - and the high is *high* (load average of over 40) - here's an example:

The current system state is: [ Load (1-minute): 67.56, Load (5-minute): 17.75, Load (15-minute): 6.23, Memory Free: 1538MB, Disk Free: 901695MB, Swap Used: 1070MB ]

Event Details:
active hosts = 15
cpu system = 0.08665106
cpu user = 0.1381733
disk free = 880G
disk free percent = 97%
disk total = 905G
disk used = 24G
disk used percent = 3%
load1 = 67
load15 = 6
load5 = 17
mem buffers = 0
mem cache = 41M
mem free = 1G
mem free percent = 39%
mem total = 3G
mem used = 2G
mem used percent = 61%
swap free = 10G
swap free percent = 91%
swap total = 11G
swap used = 1G
swap used percent = 9%
time stamp = 2020-08-04 14:03:55.279

I have no idea what causes this - and I have one host with the same hardware which whenever I access it takes a good 2-3 minutes just to draw all the dashboard widgets (and shows CPU of 10+ on the VU meter style gauge)

Any thoughts on what I should change/check?
The boxes are running full paid UT and tend to have all the major modules running (Web filter, virus, reporting, bandwidth control, etc)

I thought maybe these are DDOS style attacks - but also thought shield should take care of that?

Thanks,

Craig

[jcoffin]

I would open a support ticket since you have support. These types of issues are very difficult to diagnose without directly access to the box.