Best practise for host check icmp and smart ping

Troubleshooting
1

CMK version:Community 2.4.0P24
OS version: Ubuntu 24.04

Hi All,

I am getting a issue or question about the default host check icmp and smart ping. Currently I am using the default host check icmp for the host, but I noticed that when the ping is having packet lost or latency reaching the critical threshold, checkmk would treat this as host down rather than the ping service state change and send out the host down alert, supposed it should send out the ping service critical alert right? Should I use smart ping instead, which will explicit add the ping service under this host , and send out the service state change alert correctly ? Please advise. really confused about the default host check icmp, its description says it’s ping service but actually checkmk treat it’s a host check.

for enable the smart ping, should I do the following 2 steps? please advise

1.add the host check command as “smart ping”

image
image2332×1338 184 KB

  1. add the check hosts with ping rule under 'http,tcp, email"
    image
    image2454×1448 198 KB
2

The smart ping host check is only available in the paid editions with the Checkmk micro core.

3

but how come I tried enabled the smart ping and it seems to work ok ?

4

if the smart ping is not available in Community edition, that means the default host check command will trigger the wrong host state up/down alert in the situdation that it’s actually the service state change (ok/critical), correct ? No way to rectify this ?

5

There is a technical difference between a host check and a service check.
They both can use the same Nagios plugin.

Both rely on the return code of the Nagios plugin. If that returns != 0 for events where the host is not down, then its parameters should be changed.

But aside from that: How should a monitoring process be able to distinguish between a down host and packet loss in the network in between?

6

Hi Robert,

so can you tell me what is the best practise to setup the host / service ping check ， and what is the expected behaviour? I am thinking I should explicit add a ping service rule in “http,tcp,email..” , but I am not sure what value should set for the RTA and the packet lost rate ? For the packet lost, I think I can set the warning/critical as 80%/90%， so when the packet lost rate is for example 93%, it will trigger the service change alert and when it change to 100% rate it will trigger the host change alert. But the difficult one is the RTA , as I know the default RTA in the host check rule is 200/500, then what value should I set in the RTA in here , suppose I should set a value smaller than 200/500, right ?

is my understanding and approach is correct?
As there is situalation that , when the RTA has passed the 500ms， checkmk consider it’s down based on the host check logic, but actually it’s not down, it just RTA has passed the threashold !! It should trigger the service change alert rather than the host down alert!

If it’s for packet lost is reaching 100%, it doesn’t matter it’s host down or service down, but for the RTA reaching the critical threshold, it will confuse the receipient when he receive the down alert but he can still able to access the monitored host.

image
image1542×893 58.7 KB

7

A rule in the ruleset “Check hosts with PING (ICMP Echo Request)” creates a service check, not a host check.
AFAIK you cannot parametrize the standard ping host check.

But you can create a rule in “Host check command” that uses a custom check plug-in check_icmp which command line you can enter there including parameters.

Or use “Use the status of the service…” there and point it to the PING service check created in the ruleset above.