Packet loss Triggers with no loss

[mdiorio]

Hi Everyone,

We've been using Zabbix for about a year now and love it. We recently moved the server to a colocation facility and are monitoring things in multiple locations from this server.

Since we moved it, we have been receiving packet loss detected triggers for switches in our HQ network, which is connected to the datacenter where Zabbix lives by a 500MBps MPLS circuit. The MPLS averages 50% utilization and never over 80%.

During the time we receive the packet loss, I run a ping from the zabbix console to the switch in question. I get:
1805 packets transmitted, 1805 received, 0% packet loss, time 1806604ms
rtt min/avg/max/mdev = 1.752/4.339/37.495/4.341 ms

Doesn't seem like an actual problem, but not sure why Zabbix is flagging it as an issue.

Any thoughts? Next step for me is disabling the packet loss monitors altogether.

Thanks!

[batchenr]

Hi Everyone,

We've been using Zabbix for about a year now and love it. We recently moved the server to a colocation facility and are monitoring things in multiple locations from this server.

Since we moved it, we have been receiving packet loss detected triggers for switches in our HQ network, which is connected to the datacenter where Zabbix lives by a 500MBps MPLS circuit. The MPLS averages 50% utilization and never over 80%.

During the time we receive the packet loss, I run a ping from the zabbix console to the switch in question. I get:
1805 packets transmitted, 1805 received, 0% packet loss, time 1806604ms
rtt min/avg/max/mdev = 1.752/4.339/37.495/4.341 ms

Doesn't seem like an actual problem, but not sure why Zabbix is flagging it as an issue.

Any thoughts? Next step for me is disabling the packet loss monitors altogether.

Thanks!

can you see if you have unsupported items on that switch host?
if so- disable them all and see if it helps.
try to look at zabbix logs to see maybe you are losing connection and fot more info in General.

[mdiorio]

I did have about 5000 unsupported items. Guess that's what I get for using generic Cisco templates. They are all disabled. Within a half hour, I got a packet loss trigger.

So here's the strange thing. Whenever there is a packet loss trigger, the max loss as seen in the item graph is ALWAYS 33%. It's completely random and impacts all network devices completely randomly.

In the zabbix_server.log I grepped error and do see a bunch of these:
redundancy checksum errors.'
redundancy checksum errors.'
redundancy checksum errors.'
redundancy checksum errors.'
redundancy checksum errors.'
redundancy checksum errors.'
redundancy checksum errors.'

but not sure it's related at all, and everything else in Zabbix is working perfectly.

[batchenr]

I did have about 5000 unsupported items. Guess that's what I get for using generic Cisco templates. They are all disabled. Within a half hour, I got a packet loss trigger.

So here's the strange thing. Whenever there is a packet loss trigger, the max loss as seen in the item graph is ALWAYS 33%. It's completely random and impacts all network devices completely randomly.

In the zabbix_server.log I grepped error and do see a bunch of these:
redundancy checksum errors.'
redundancy checksum errors.'
redundancy checksum errors.'
redundancy checksum errors.'
redundancy checksum errors.'
redundancy checksum errors.'
redundancy checksum errors.'

but not sure it's related at all, and everything else in Zabbix is working perfectly.

i think this error comes from one of your cisco template
if you add this switch with the regular templates like Template SNMP Generic
can you see how it works for you ?

unlink & clear all templates first.