Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials

Scalability of NPM

This might be backwards but I've recommended Orion to my manager and he agreed that we should purchase it, and only now am I running the demo. I'd first like to say that all of you who said that Orion was easy and quick to setup are smoking crack!

Seriously, it may be easy if you don't want to customize anything, set up any alerts (with appropriate suppressions), create maps, etc. I'm really beginning to doubt my recommendation because I am foreseeing an immense amount of work.

I'm going to have to load in all of our devices, create custom properties, create maps, figure out some sort of suppression scheme that makes sense, create alerts (a LOT of similar alerts if I want to have sensible suppression), etc.

I'm not even sure what I want out NOC screen to look like. Maps are really going out of favor. Seriously, how often have you had an icon on a map turn red and stay that way for an hour before someone notices? I thought about having email alerts for all Node Down events but then what happens if our main site loses connectivity with the rest of the network? Our mailboxes will have hundreds of alerts in them because Orion will think the entire network went down. So, email alerts might be out unless I can figure out how to design suppression without having to create 1000 different alerts.

In the end, I think the customization will pay off but it doesn't seem to be quite as easy as every says. At least if you're doing it for the first time, that is. I'm sure it gets much easier once you figure out what you're doing!

Anyway, to the point of this post. Do any of you have any thoughts on how to scale Orion so that it is still easy to use, it provides a nice NOC status screen, and can send out alerts in a manageable manner?

Thanks!
John

Find more posts tagged with

Accepted answers

All comments

Network_Guru

quote:Originally posted by jneiberger

I'd first like to say that all of you who said that Orion was easy and quick to setup are smoking crack! :-)

Anyway, to the point of this post. Do any of you have any thoughts on how to scale Orion so that it is still easy to use, it provides a nice NOC status screen, and can send out alerts in a manageable manner?

Thanks!
John

Thanks for the kind words John :-p

I'm not sure what other NM products you have setup, but nobody said it was going to be quick, just easy.
Depending on the size of your network, yes, this can be a huge task.
I believe the point made, was that using the autodiscovery feature you can get meaningfull data from your network devices in a few hours. Customizing that data is what takes time, but is also the strength of the Orion product. Most products have fixed views & only allow you to present the data in a pre-determined format. There are Orion users on this forum who have spent considerable time writing their own code to make Orion suite their needs. Not many products can do this.

As for initial configuration, I would start out with the default views & work from there.
I have not found the Map view terribly useful, (lots of work & maintenance for not much return in info) but nice to look at if your are a CIO or VP

My home page consists of 2 columns:


Down Nodes Nodes with High Network Utilization
Search for Nodes Nodes with High CPU Utilization
Nodes with Packet Loss Events Today
Nodes with Hi Latency Syslog events
Search for Interfaces Top 40 Reports (Report Writer reports)
Nodes with Problems (Interface errors, etc.)

This shows me my overall network status in a single page.

As for alerting, I think that everyone has agreed that the current non-beta (native) alert engine is terribly painful to setup. If I was doing it all over for the first time, I would definitely use the new Beta engine, but only after I have setup Custom properties for all my nodes with Custom Property editor. That will make creating alerts & suppression MUCH easier. Base your alerts on these Custom properties, not the SNMP discovered inforamtion. Yes this is time consuming, but you only have to do it once, and every time you add a new device to be monitored.

Happy customizing!

-=Cheers=-
NG

pblatz

quote:Originally posted by jneiberger ...
Anyway, to the point of this post. Do any of you have any thoughts on how to scale Orion so that it is still easy to use, it provides a nice NOC status screen, and can send out alerts in a manageable manner? Thanks! John

John, What product do you currently use to show the NOC the current network status? We did something very similar to NetworkGuru as far as our main page goes. I started by manually adding our internet routers, external and internal firewals and core switch/routers. I then looked at all the pre-defined custom fields and added those that most suited our needs. Next I added 5 additional fields

Esc1
Esc2
Esc3
PriContact
SecContact

These fields define my escalations and notification groups. Esc1, 2, 3 are True/False Fields Pri-&SecContact are my first and second level support groups. I use the included alert engine. I send all Node Down alerts to NetCool as a SNMP Trap. I modified the trap sent to include the 5 custom fields which are evaluated by Netcool to determine severity/criticality. I also pass the "IsRouter" attribute. Additionally all our devices have a preset naming convention such as: AAccccNxx## where AA == Country Code cccc == Company specified SiteCode based upon CityAirport code and site type; for example, HOUS = Houston N == Network Device xx == Device type == rt = Router, sw = Switch, fw = Firewall ## == Sequence number. These are just examples, I recommend you come up with your own naming convention and publish the standard for you company. We change the N to other letters for other devices (L=laptop, S=Server (role is defined in the xx field), D=Desktop, P=Printer) Additionally I suggest that all devices be included in DNS and you are able to resolve forwards and backwards. As far as Maps go. We have a visio file of each logical network and I have a link on our main page which opens the shared filestore for all our drawings. We are planning on publishing them as webpages and then will provide the links to the apropriate visio page. Also, I am investigating what it might take to create a "dynamic" visio map by querring the Solarwinds database. Thanks, P. B. Success requires three bones--wishbone, backbone and funnybone. Kobi Yamada

DannyMc

quote:Originally posted by jneiberger

I'd first like to say that all of you who said that Orion was easy and quick to setup are smoking crack! Seriously, it may be easy if you don't want to customize anything, set up any alerts (with appropriate suppressions), create maps, etc. I'm really beginning to doubt my recommendation because I am foreseeing an immense amount of work.

I am one of those that said Orion was easy and quick to setup. I had it up and running in 45 minutes. I did the discovery and only selected the devices that I wanted to monitor (infrastructure). I had e-mail alerts setup and configured in 10 minutes. I had Active Directory integration setup in 10 minutes. I fumbled around with the map tool -- of course I did not read the instruaction -- and had a simple map working in 2 hours. I spent most of this time looking at all the canned map images.

You said the "C" word. I learned long ago that "custom___" and "easy" are a contradiction in terms.

Just to implement basic collection, alerting and a map I was up and running in an afternoon. I came from OpenView and CiscoWorks. Speaking of smoking crack, try doing "customization" in those!

I'm going to wait for v8 for the new alert engine to come out. Add a modem to the server and set up an alert monitor place an old fashion voice call to your cell/home phone via the modem if the core or e-mail server are down. Play a voice recording over the dialed phone call to alert you of the condition. Loop it so that if it rolls to an answering service you can at least get notification there.

Years and years ago we used to have a room monitoring system that would do such a thing. I think it was from Unisys. It would call up a predefined phone number and repeat over and over the alert (ie: temperature, water, etc.). Same concept here.