Church Crosstalk goes with ActiveXperts Network Monitor

Authors

Editors of ChurchCrossTalk.com. ChurchCrossTalk.com is a forum for churches to help churches to share their experiences in the area of information technology.

Date

Article was released on 2004/06/28.

Abstract

Does anything bug you more than one of your users telling you that one of your servers is down? Perhaps it's just me, but I feel like such a failure when that happens. Here at CCV we now have 8 servers, 12 switches and several other devices that on a rare occasion decide to take a break from their normal duties. With the addition of our new worship center it's time for use to take a more proactive strategy our system administration. With that in mind I set out on a mission to find a solution. Below is some of what I learned and a few product suggestions.

On Monitoring and Metrics

Let me first tell you what I was looking for. I wanted to find a product that could a: tell me when a server, switch, or other device was not responding or had a critical event (defined as monitoring) and b: would be able to display performance metrics over time (defined as metrics). Oh yeah, and it better be cheap! No IBM Tivoli or HP OpenView several thousand dollar apps got it!

After reviewing and demoing over 25 products, I found that there really was no one package in my price range that was able to fulfill both needs. Trust me, this wasn't an easy conclusion to come to. I feel as though I chased a thousand rainbows, only for them to disappear in front of me. In the end I was able to find two separate products that together gave me what I was looking for. I address each below.

Monitoring

Of the two requirements this was the most important. I want to know when critical events happen before my users do. Some of the requirements that were important to me are bulleted below.

Must provided basic outage detection (i.e. does not respond to a ping);
Must be able to detect performance issues on Windows servers (i.e. high CPU utilization, low disk space, low memory);
Must be able to intelligently determine in network services like HTTP, SQL Server, and FTP are not responding correctly;
Must be able to interface with SNMP (a standard monitoring interface);
Must be able to send not only e-mail notifications but also pager notifications via a modem (e-mails fine until it's the e-mail server that goes down);
It would be nice if it had a web interface to check the status of items remotely.

After pulling my hair out and testing every product under the sun (well I'm sure I missed one or two) I decided to go with ActiveXperts' Network Monitor Manager. It may not have the fancy network maps the market leaders do, but it works well and is simple to use. I'm actually disappointed that the market leaders, NetCrunch and What's Up Gold, made it so hard to do simple things like sending a notification based on a SNMP trigger or Windows performance event. With ActiveXperts this is a snap. I had the whole thing figure out in 5 minutes yet was able to do advanced notifications. OK… OK… I hear you, what's the price. They offer a 50% discount to non-profits so you can get an Enterprise license for $350. This is a real steal for what you get. If you still think that's high you can actually use the product for free (yes FREE) with a limited feature set (basically it will only notify you when your server stops responding to pings or it's HTTP service dies, basic but still usable).

We're currently using it to monitor the following:

Servers:

Ping
CPU Utilization on 90%
Disk space on all drives
HTTP
SQL Server
Terminal Services

Switches:

Ping
Telnet

Other Devices:

Temperature of our MDF and particular IDFs. We have one IDF closet without dedicated A/C. We have it page us if the temperature goes over 90 degress.
Cisco Wireless Access Point (currently we just ping it to make sure it's up, in the future we hope to use more advanced SNMP monitoring)

There are many other uses we hope to try in the future. One interesting one would be to have it notify us of printer jams.

Metrics

The other half of the problem is metrics. I want to know the load on my servers so that I can proactively look at required upgrades (especially disk). The product I found that best meet my need was PRTG from Paessler. It's actually a simplified version of MRTG the fames Linux/Unix SNMP graphing utility. When I say simplified I mean it… it's real easy to use if you have a basic understanding of SNMP. If you don't I wrote up a quick overview on using SNMP in a windows environment. You can download the Professional version for $50. Now that's a deal (you could download MRTG for windows for free, but you'll save the $50 in labor and frustration in the first week).

The premise of this product is simple. You provide it with a SNMP OID (translation a metric like CPU utilization) and it will make a nice graph and post it to a webpage. The graphs may not impress Ross Perot, but they have a technical elegance to them. Some of the metrics we're collecting are listed below.

Servers:

CPU Utilization
Available Memory
Available Disk - Very Useful
Number of Active Terminal Server Sessions

Firewall:

Bandwidth Utilization

Misc:

MDF and particular IDF room temperatures.

Summary

Now that both products are in we are learning some interesting things about our environment (especially the performance our servers and how often the AC turns on in our IDF… OK that last one may not be that interesting). I certainly hope that this has been helpful to a few of you. This project took a lot of my time; I hope that this article will save you some.