Solarwinds

[Fred]

I'm deploying SolarWinds at work.  My background is in open source cacti/nagios/etc., so this is somewhat new to me.  Experience so far is that it does way more out of the box, and then I hit a brick wall when I want to do something custom, but anyhow...

Anybody have any tips, tricks, or general advice for using Solarwinds for network monitoring?  We currently have NPM, NCM, NTA, UDT, and VQNM.  We separate SAM into a separate deployment because servers are managed by a different team, but there are some features I wish I had.

[Ironman]

I haven't had a chance to dig into NPM and NTA but it seems like they have some really neat features.

We mainly use SolarWinds for the NCM piece and jobs! Which are awesome!

[deanwebb]

Netflow monitoring on SolarWinds isn't as comprehensive as on a security platform, but for seeing all the traffic flows and getting that kind of visibility, it's great stuff.

As with all monitoring tools, it's as good as the gardener that tends it, and how regularly he tends it. A beautiful setup for today is going to be a mess a year from now without maintenance.

[Seittit]

I have extensive experience with the Solarwinds suite of products, and currently own one of the largest implementations of NPM their support and sales team has ever seen. Every time I open a support case, their engineers make this face on the phone: 


The product is very stable, but always look up release notes to find resolved caveats and outstanding bugs. Solarwinds scales quite nicely, you can just add additional pollers if you need to monitor additional "elements" (interfaces, volumes, etc.)

The alerting features are top notch, and highly customizable. They have great support and their Thwack forums are very helpful too. We also do some clever additional things, such as using the SWIS API to redirect event logs to HP BSM, which then takes the alerts and forwards them to X-Matters alerting system.

To top it all off, they're based in Austin. Gotta show love for Texas

[that1guy15]

Agree the forums are a great resource. Documentation is good to. Just spend the time and run through it.

Most of my deployments have been Nagios/Cacti or Zenoss and I agree customization and tweeks are simpler in them but they are there in solarwinds. Just like any NMS there is a learning curve.

Right now Im in the middle of a Solarwinds deployment done wrong and it might as well be set on fire and shot in the head its that useless. But not my monkeys and not my circus!

[scottsee]

Interesting thread.. Our 2nd quarter budget was approved which includes Solarwinds products. I've been reviewing their network, hardware, event, Sql and email archiving products for implementation. We currently use LinkMonitor on the cloud.. Costs us roughly $1000 a month in subscription fees..

[Fred]

Agree the forums are a great resource.

Odd... I find Thwack to be a cluster of poorly organized information, and the people there to rarely be helpful.  Trying to find current documentation/sound advice on how to back up my F5's with NCM has led me to a half dozen half implemented solutions. And I'm all confused about where maps are and where they are going... I don't think I know my terminology on those yet.

I'd much prefer standard forums instead of what they did, which seems to be to reinvent the support forums wheel and to do it poorly.

Right now Im in the middle of a Solarwinds deployment done wrong and it might as well be set on fire and shot in the head its that useless.

Can you talk about a couple of the particular things they did wrong?

[that1guy15]

Its a centralized deployment at HQ with only one site w/ a remote poller. They spec'ed the centralized server too small so we are limited on what we can do. Once NCM was setup I deployed to all my network gear and within an hour brought down the whole server. So I had to back NCM off to a handful of devices until they figure out what to do.

This is more layer 8 issues, but since the Solarwinds deployment is centralized at HQ the team there setup all the page layouts and templates to what they care to see. Since they dont care about all the details at my site those things were axed form the templates. This leaves me with useless information for all my devices. Also since I am limited on what pollers Im allowed to deploy I cant get proper alerting and failure notifications on things like PSUs, Fans, Route flaps and STP issues.

The organization of all the devices and sites is not properly deployed so filtering and managing Syslog/Traps is impossible. Nor has anyone took the time to straighten it out. So the best you get for those is stare at a streaming wall of text and hopefully see something... Luckly I drop those locally to another server as well in case I ever need them.

This deployment is the classic NMS problem most companies face. At some point leadership decides they need proper monitoring and figures just a quick PO to purchase a product will solve all their problems. Fire and forget the server and all is good with the world. Usually the task id dump to some Jr engineer with little knowledge of NMS and boom, your  traditional half-baked deployment.