The dashboard shows the total usage across all cores. So on the 2100 that's 2 cores.
That could be both cores at 90% or one core at 100% and the other at 80%.
Diag > System Activity should show you what's using it. 90% is very high unless it's passing a load of traffic. Running top -aSH at the CLI shows the same thing
On a test box here I see:
last pid: 72786; load averages: 0.17, 0.32, 0.26 up 0+02:28:52 15:38:17
623 threads: 3 running, 595 sleeping, 25 waiting
CPU: 1.0% user, 0.0% nice, 1.8% system, 0.6% interrupt, 96.7% idle
Mem: 82M Active, 683M Inact, 643M Wired, 1902M Free
ARC: 428M Total, 123M MFU, 282M MRU, 552K Anon, 3223K Header, 19M Other
360M Compressed, 743M Uncompressed, 2.06:1 Ratio
PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU COMMAND
11 root 155 ki31 0B 32K CPU1 1 136:03 96.78% [idle{idle: cpu1}]
11 root 155 ki31 0B 32K RUN 0 137:38 96.08% [idle{idle: cpu0}]
21079 root 25 0 139M 44M accept 0 0:10 1.39% php-fpm: pool nginx (php-fpm){php-fpm}
12 root -60 - 0B 400K WAIT 0 1:13 0.82% [intr{swi4: clock (0)}]
72404 root 20 0 14M 4788K CPU0 0 0:00 0.72% top -aSH
0 root -92 - 0B 448K - 1 0:47 0.42% [kernel{dummynet}]
23993 root 20 0 28M 8084K kqread 1 0:01 0.22% nginx: worker process (nginx)
12 root -92 - 0B 400K WAIT 1 0:11 0.12% [intr{pcib0,0: ath0}]
78950 root 20 0 11M 2708K select 1 0:09 0.10% /usr/sbin/syslogd -s -c -c -l /var/dhcpd/var/run/log -l /
89921 root 20 0 405M 361M bpf 1 0:39 0.09% /usr/local/bin/snort -R _28847 -D -q --suppress-config-lo
12 root -92 - 0B 400K WAIT 0 0:05 0.07% [intr{gic0,s42: mvneta0}]
57454 root 20 0 15M 2624K nanslp 0 0:04 0.05% /usr/local/bin/dpinger -S -r 0 -i GIF_TUNNELV4 -B 10.2.4.
9 root -16 - 0B 16K pftm 1 0:09 0.05% [pf purge]
Steve