NetXMS Support Forum

Hello,

I am trying to create an alarm that will trigger wen the outbound traffic on some ports reaches 50%,75% and 100%, but since my knowledge of the program is very little I would like some help.

Thanks.

What device type? Windows? Linux? Switch/Router/Firewall?

The basic idea is that you will need to get the bandwidth for the interface. Which can be a hard coded value or perhaps gathered with Data Collection. You will then need a DCI to get the outbound traffic. In the Transformation section of the outbound traffic DCI put

(($1*8)/GetDCIValue($node,9363))*100

Where $1 is the current value gathered by the DCI, GetDCIValue is the DCI you created to gather the current bandwidth. $node means, you are looking at the current node for the DCI ID and the 9363 is an example of what the ID would for the bandwidth DCI. Or instead of GetDCIValue you could hard code it. A 1Gbps interface would be 1000000000.

To create alarms, use the Threshold section of the outbound traffic DCI. To configure events, use the Configuration/Event Configuration option from the main console. You will want to do that before you create the thresholds so the events will be available for selection in the Threshold

I hope that gets you started. There also could be other (better?) ways to do this. NetXMS is very configurable. If you have more questions, please ask and I will do my best to respond.

Thx gdodd I am hard coding and using thresholds to get the alarms wen it get to 50% and 100% of data transfer, I am now going to make the calculations to get the 75% alarm.
The problem that i have now is that yes it gives the alarm wen it goes hover the 50% mark but i want the notification of the alarm to stay to see at what time it happened.

It sounds like you have the same thinking with alarms as I did when first using NetXMS. If you want to know what time the alarm happened, you look in the event log. I used to keep all my alarms showing so i knew at what time something happened. But that quickly became cluttered. I now terminate all my alarms when the issue is resolved and check the event log for what time things happen. I think of alarms as the current state of the node and event log gives me the history of alarms. An easy way to see events for a node is to right click the node and choose Event Log.

If you want the alarm to stay. You need to configure Event Processing Policy from the Configuration menu. Sounds like there is a policy to terminate the alarm when the deactivation event is happening. Instead of terminate you could use resolve for the policy. That will leave the alarm, but show the node as healthy (not in alarm) and leave the alarm for manual review/termination.

Thx again, in that case I don't need the alarms to stay since I can just go to the event log to know wen it happened and all the information regarding that moment.