Check_MK How to change thresholds for RTA and Packet loss

[Gustavo]

CMK version: 2.0.0
OS version: Centos 7

I have a question regarding the default thresholds that CMK use.

I see CMK use a standard for the RTA and Packet loss configurations (RTA 200ms/500ms Packet loos 80%/100%), you can see these thresholds clicking on the name of the host:

In this section, when i click on the host icons part to go to the “Parameters for this host”, i can´t create a new rule to modify the current thresholds, i only can see the main configurations for this host like other kind of rules, but i don´t know if i can change this thresholds:

Does anyone know in wich part of this sections i can change the default thresholds for the RTA and Packet loos?

Even if i activate the PING service check, and configure another thresholds for the rta and packet loos, cmk use both parameters and i don´t want this PING check activated, i just want use the standard configuration.

[lkoenig (Lorenz Koenig)]

Hello @Gustavo,

try the following rule:

This should have the desired effect.
However, this may be different in the Enterprise versions, but i assume from your tags for this topic that you are using the RAW edition.
Hope that helps.

[Gustavo]

That’s exactly that i was looking, thank you for the information Sir

[LaUs3r]

Hi, I am struggling with adapting the RTA of the PING service. I have a solar inverter which somehow shows in check_mk crazy roundtrip times (in the screenshot below 1.294.133ms!!!).

For testing, I put the solar inverter into the networking segment “WAN (high latency)” and changed the host check rule accordingly:
-w 7540000.00,80% -c 8540000.00,100%

As you can see, those thresholds are not applied by the rule - it seems that 255000ms is the maximum that can be used:

I use checkmk Raw Edition 2.1.0p24.

Is there any hint you can provide?
Thx

[gstolz (Gerd Stolz)]

hi @LaUs3r,

• next time, please open a new topic this one has been solved -

I guess it’s possible that there is a maximum value that is set as the threshold, what I’m wondering is: maybe there is a tcp port you could check (web interface or something like that?) that responds better than the ping?

Gerd

[LaUs3r]

@gstolz, yes, you a right. creating a new thread would have been the better choice.

Nevertheless, I’ll check the open ports. Maybe that is more successful. Good idea (I could have though of by myself )

[LaUs3r]

Just to provide a final feedback: the hint with the TCP check of a dedicated port did the trick! It’s all working now.
Thx @gstolz