Zabbix queue length when adding ping check on host

[Gijs007]

I've added a ping/icmp echo check, which pings an IP from the host which is monitored.
The goal is to measure the latency and the packet loss.

For example:
PingLatency[8.8.8.8]
PingLoss[8.8.8.8]

UserrParameter in the Zabbix.conf file:

UserParameter=PingLatency[*],ping -w 2 "$1"
UserParameter=PingLoss[*],ping -w 2 "$1"


The issue I'm having is that the Zabbix queue length increases when I add these items as active check. I believe the reason for this is that the check takes a few seconds to complete, as it's sending four pings. During this period everything else needs to wait, as the active checks do not run in parallel.
When using the passive check method the results don't come in every minute. Even though the check is configured with a one minute interval. I believe this happens, because I run many of these checks (200-500 passive checks every minute).

I'm looking for any advice to resolve this. The goal is to measure the latency and packet loss from the monitored host to a destination specified in the Zabbix server. While getting regular (e.g. 1 minute) updates, and not causing any queues in Zabbix.

[gert.derouck]

You could install a zabbix proxy on that host, that initiates the pings. And add the devices that you ping, as normal hosts, managed by that proxy.