One of my client is asking if we can use DC-RUM to monitor the switch and port, i.e. if we can identify which switch/which port has problem if the application is not running ok because of network problem?
I have check the below two links about netflow:
It seems like there is no filed from the netflow data to identify the switch id? And in the second link, there is one scenario to use network collector to identify the switch problem, anyone can help explain more on how dc can do this?
DC RUM is not designed to be a Network device monitoring platform (up/down of ports, host metrics, etc), we more or less care about the 'flows' that are going through that device so we can tell what kind of traffic is consuming the link.
Saying that, if you were receiving Netflow from a router and you were receiving flows from the WAN interface. DC RUM could tell you what traffic is passing through and the volume metrics associated with those. It will also pull back the interface type (serial, fast ethernet, gigabit, etc.) and port speed (what speed has been agreed for that link). Knowing this information you could solve the following use cases:
- Who's eating all my bandwidth?
- Do I have high utilisation coming out of that interface?
- Is that interface down (you would know because no traffic would be flowing so your throughput graphs would zero out)?
In addition, you could also configure alerting to notify for things like high/low utilisation. This can be done in the RUM Console in the Alerts sections just like the other alerts.
Because of the nature of Netflow and the amount of data being passed to the AMD. I wouldn't recommend turning on Netflow on the core switch or any LAN device for that matter as it will most likely overwhelm the AMD and/or it is an inefficient way (overhead, duplicates) to derive an answer. By enabling it on the edge device (router), you'll know everything that is going in/out of the site, your flow volumes will stay under control, and you'll solve majority of the typical use cases.
Hope that helps!