Notification linked to a group

Dear Laurent Piette,

Thank you very much for your KB-post.

Just to make sure, this is what I think your structure looks like (logically):

[d]Router 
|__[g]Rooms (depends on the ping of the router)
     |___[d]Room 1 [[Control Panel]] [Audio] [Temperature]
     |___[d]Room 2 [[Control Panel]] [Audio] [Temperature]
     |___[d]Room 3 [[Control Panel]] [Audio] [Temperature]

Legend: 
[[name]] = master object for the Room device
[g] = group
[d] = device

Is that about right? If not, please post the complete structure, so we can help you configure dependencies correctly.

Using dependencies, you can pause sensor monitoring based on the status of another sensor to avoid false alarms and incorrect downtime recording. A dependency stops the monitoring of one sensor or a set of sensors as soon as a specific sensor is in Down status.

This means, for example, that you can stop monitoring your audio systems when the corresponding switch is down due to connection problems.

In this way you can avoid notification flooding - instead of notifications for each sensor on that location, you will receive only an alert that a master sensor is down and all other sensors will be paused.

If you want to set a sensor to be the "Master Object" for its parent device, please click on "Settings" of that sensor, scroll down to "Schedules, Dependencies, and Maintenance Window" and under "Dependency Type" please choose "Master object for parent" and click on "Save".

You can also set the "Master Object" for a device/group in the following way: click on the "Settings" of that device or a group, scroll down to "Schedules, Dependencies, and Maintenance Window" and under "Dependency Type" you can choose which sensor will be a master object for that device/group.

To see the map of all already existing dependencies set by a former employee, click on "Devices >> Dependencies >> Dependencies Graph" and there click on "Show Master Dependencies".

If a Ping sensor is set as a Master object for its parent device, and if that Ping sensor is down, then all sensors on that device will be paused. However, if a HTTP sensor on that device goes down, the Ping sensor will not be paused, because the HTTP sensor is not a Master object for the device.

For more details, please see:

• https://www.paessler.com/manuals/prtg/dependencies
• https://www.paessler.com/manuals/prtg/device_settings#schedules
  

Once you have done this you can set a notification on the master object of each room, to get noticed when the room goes down.

Kind regards,
Sven

Hello Sven,

Sorry for the delay in getting back to you. Here is the structure.

[d]Router
|__ [g]Room 1 
       |__[d]Control panel (Device always UP)
       |__[d]Room controller (If up, room is up, if down, room is down - depends on the ping of the Control panel)
       |__[g]Room group (depends on the ping of the Room controller)
             |___[d]Device 1 [Video]
             |___[d]Device 2 [Audio]
             |___[d]Device 3 [Temperature]

I already have some monitoring, Control panel, Room controller which gives the status of the room (UP and Down). These are working normally taking into account all dependencies. I've also added a Room group monitoring for all the Device into the group.

The problem is that everyday, the room are switched off and I'm getting alarms for Devices down. I'll check what is the switching off order, it should be the controller first.

Regards

Dear Laurent,

you could set the ping sensor on each device to "Master for parent", and also enable Auto-Acknowledge for the sensor (both settings available in the Ping sensor's setting tab.)

When the device is going down, the ping sensor will fail, auto-acknowledge itself and set its parent, meaning the entire device object, to "Pause".

In order to make this work correctly, use a 30-seconds scanning interval for the ping sensor and a 60-second delay for the notification state triggers.

Hello,

Devices in the room have Room controller ICMP as parent. Room controller is having Control panel ICMP as parent.

ICMP is currently "Show down status on error 'default''.

When the Control panel is triggering a shutdown, all the devices in the room including the control panel is going down but I always get a few off them before the room is going to "Pause".

The Control panel can send a SNMP trap before starting the process. Can I use it to pause the room?

Regards, Laurent

Dear Laurent,

pausing with traps is difficult. In principle you can do it: Let the trap receiver trigger a notification which performs an API call with sets the master sensor to "simulated error".

Similar, another trap receiver sensor could do a notification to resume. However this is complex in setting it up. It would be easier to check that the notifications for the sensors which should be paused are triggered with a delay, so that they are not triggered until the time, the ping sensor pauses the device(s).

The sensors might stell turn red, but at least trigger no notification anymore. If could be necessary to remove the inheritance property of notifications from the object parents and create new notification triggers with a longer delay.