For Example : I have to monitor a switch or other network devices but i don't have the list of parameters so, How can i know which parameter needs to be monitored. May be it could be client Decisions but skyline have any reference list for that?
Hi Alberto,
Cisco Switch
I see, sharing my experience below.
There are several strategies you can follow - normally you'd need some input from the system owners (e.g. if a set of ports is known to be unwired / admin down, and if they want to ignore them in the monitoring criteria or if they want to be alerted when their state changes).
A first iteration could establish what are the relevant KPIs, which ones need alarming, which ones need trending - the following iterations will refine the monitoring criteria until you have a set of rules that meet your DoD "definition of done" and you can roll out in the production environment.
Other approaches also work, however multiple iterations (between DM admins and system owners) will allow you to have less false alarms in your environment. Once it's done for one element, it's releatively quick and easy to extend to all the other elements of the same type in your DMS (I'm saying type, not protocol, as a LEAF node would possibly have a different set of KPIs when compared to a SPINE, even if they are polled with the same DataMiner protocol, hence the importance of dealing with the system owners to establish your monitoring criteria).
So, I need to check with system owners or manufactures. As Well as if already created drivers against elements i can refer that right?
Yes, there are several parmaters you’ll find defined in a DM protocol, especially in complex ones for network equipment: bandwidth, discarded packets, state of the ports… are essentially something that could be monitored, but to what extent can be determined only with a good knowledge of your service architecture.
If there is protocol already built, I’d start from the parameters listed for the Alarm template and Trend Template – is this CISCO? Juniper?