Major Zabbix has been using ~100% of my cpu on two different servers.
The common thing between them is that they are both being contacted by another server that collects the data.
Its zabbix-server process died because of high cpu load (ddos attack), and I think it triggered this behaviour.
zabbix on the web frontend version 1:2.4.8-1+wheezy
two high cpu servers zabbix version 1:2.4.8-1+jessie
both have these packages installed from zabbix repo
zabbix-agent
zabbix-frontend-php
zabbix-server-mysql
# ps aux | grep zabbix
zabbix 8040 99.3 0.2 27824 5520 ? R May28 1577:44
# strace -p 8040
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
select(8, [6], NULL, NULL, {0, 0}) = 0 (Timeout)
^Cselect(8, [6], NULL, NULL, {0, 0}Process 8040 detached
<detached ...>
please let me know if any more data is needed