Monitoring NSX-T

[mackov83]

We are about to start the migration from NSX-V to NSX-T. All the testing went ok, however at the point of monitoring and alerting, the product seems to be very underwhelming. 

While NSX-T seems to support SNMP, the VMware support team advised that only traps will work, even though the configuration for SNMP via the GUI clearly lists both traps and polling.

The following are options I have in mind, though each has its pros and cons:

• SNMP - traps only
• Log Insight - alerting only. Though seems more useful than SNMP traps
• vRealize Network Insight - so far has the best alerts and monitoring but would require additional investment
• NSX-T API - not a lot of useful information or blogs from what I can find

Interested to understand from the community what others have used to successfully monitor NSX-T. We are currently using 3.2.1

[D_G_Tal]

hi ,

BIG MON could be good but needs money 😄

[ptarnawski]

Use Zabbix, 

[mackov83]

@ptarnawski how does using Zabbix help? i.e. Have you got any useful information specifically for Zabbix monitoring NSX-T? Have you done it? I don't see anything on your blog about Zabbix. Is this just a plug for your site?

We already use PRTG for monitoring and there are plenty of monitoring tools out there. I am interested in how people are collecting the data from NSX-T and feeding that into monitoring and alerting tools.

FWIW, I tried to use Zabbix some time back (version 4.x) and found it to be one of the most complicated and time consuming products I had ever used. I spent over a week on it in whatever spare time I had and parked it.

[ptarnawski]

@mackov83 WOW, just, wow! I use Zabbix, I feed information from NSX-T to it,  however, your aggressive tone of reply and basically shouting at me made me uncomfortable to give you any further information and continue the discussion. 

You could just ask something like this:"Could you elaborate on Zabbix?"  instead you just pour your frustration/aggression. 

Thanks for that. Have a lovely day.

[mackov83]

@ptarnawski I didn't mean that to be aggressive and didn't realise the first sentence was in bold either. If you took it that way I sincerely apologise. I also didn't realise that the links to your site and channel were a signature either - a line separator may be a good idea?

I'm not so much interested in the monitoring tool which is why I originally didn't include what we use - everyone I am sure will have something different. I am interested to understand how people get the data out of NSX-T and into a monitoring tool (i.e. the method). Having asked VMware their answer was vRNI or NAP.

As I explained to VMware, NSX-V had an individual VM for each ESG, so monitoring throughput is pretty easy via the vmnic. However, each NSX-T  T0 has multiple virtual interfaces on the same vmnic for multiple customers, therefore monitoring throughput is not so easy. As we will be sharing T0s between customers, we want to avoid the dreaded noisy neighbor issue by being able to place new customers on Edge Transport Nodes that have available bandwidth.

As of yesterday, I was also made aware of NSX Application Platform (NAP) which was introduced in 3.2.x to provide some of the monitoring capabilities that existed in NSX-V - Getting Started with the NSX Application Platform (vmware.com). However, from what I can tell this involves deploying Tanzu and I don't really have the time for that right now. Likely we will progress with vRNI and evaluate further later on.

Two requirements:

1) Alerting - issue notifications

2) Monitoring - performance statistics

[dazzlr]

Wow such as shame this conversation stalled.

@ptarnawski I don't suppose I could coax you into continuing the thread with your experiences around monitoring NSX-T with Zabbix? 

In addition to the use case mackov83 mentioned, I would really like to monitor user password expiry dates. I wonder if this can be done with SNMP?

🤞