What's the current state of QoS queue monitoring? I just took a look through the old posts on the topic. One said that it was due out summer of this year.
My customer needs a way to monitor the Cisco CBQOSMIB for queue depth, bandwidth utilization per queue, and queue drops. We have NPM on order and I'm wondering if we need to order something else to monitor this MIB. A custom poller isn't very feasible due to the dynamic nature of the MIB tables and the correlations between variables that must be done to make sense of the data.
I would also like to know more about CBQOS
Can you share any document or information that would be helpful?
I want to monitor a particular dscp value and in which queue it is priotorised
Any help is appreciated
It is important to be able to know if a given traffic class is dropping packets. When congestion is happening, the low priority queue may be dropping (that's a good thing) while the high priority queues are still being handled. Knowing that the low priority queue is dropping and how much it is dropping is important for early visibility into congestion and the need to plan for a link upgrade (or other action).
I use NTA version 3.6 and find it very useful in keeping track of customer QoS policies.
We track the shaping policy on the WAN interface which can be pulled from the router via SNMP then use the flow analysis to pull which DSCP classes are in the exported traffic (and indeed imported).
In many respects, depending the policy and queue type you'll get packets dropped by the router before egress so they won't be on the flow export but it does mean you can see how much of each policy is making it through the post shaping into the transmit queue and compare that to the set policy.
You can even go one step further and pull the "show policy-map interface (x)" tables from the router via SNMP to pickup discards etc for even further enrichment but its not as pretty.
I'm really happy with it. We can from MPLS networks each nodes and the flows of DSCP.
Thanks forr the info..I`m interested in the shaping aspects of it, sounds like it will do what we need. We currently run 3.1 I`m assuming we`ll need to goto NPM9.5 before we can upgrade ?
There is a document called "Upgrade instructions" in the Customer Portal under the DOWNLOAD YOUR SOFTWARE or REGISTER YOUR SOFTWARE sections. Look in one of these areas for Orion or an Orion module areas and you'll see a link to the document set. make sure you -
The documents contain all the links to the bits for any intermediate updates you will have to make.
The CBWFQ reporting in NTA is great, and is a big help. Is it possible though to create an alert when a threshold is crossed for traffic in a particular queue? For example, with our WAN carrier, we subscribe to a certain amount of EF tagged traffic, say 3M. We want to be proactively notified via alert when traffic approaches that limit, say 80 or 90%- not necessarily that it has been reached and drops are occurring.
Sr Network Analyst
(1) Orion v10.2 SLX polling engine & Additional Web Server 35500 elements
(5) Orion v10.2 SLX polling engine
(1) Orion 7.0 NCM DCX
(1) Orion 2.0 UDT UCX
(3) Orion 3.8 NTA NCX 1500 interfaces
(3) Orion 3.5.1 IPSLA SLAX
On a side note: NTA needs to work like the new NCM. You should be able to buy just the number of interfaces you need to monitor netflow on. Its so resource intensive (and costly), we just have a 100 interface license and add devices we need to monitor. In its current configuration, we have a 100 interface NPM license to run it on and its totally independent to our production NPM.
That said, the CBQOS, Application, and TOS reports are used on any congested sites, or when VOIP or video are killing things and we need to verify ACLs and Tagging are working.
SolarWinds solutions are rooted in our deep connection to our user base in the THWACK® online community. More than 150,000 members are here to solve problems, share technology and best practices, and directly contribute to our product development process.