On even the best servers, memory goes bad, disks fail or fill up, and applications crash. Knowing what's going wrong before it goes too far wrong can make the difference between being a hero and a goat. Monitoring software can give you the warning you need to turn a production disaster into a bit of just-in-time maintenance.

Example: If a drive fails in one of your servers, the chances are good that you won't notice it. You've almost certainly configured them with RAID arrays, protecting them against the failure of a single drive. At worst the server slows down, having to work a bit harder to compensate for the missing member. All you need to do is replace the dead drive and repair the array.

Simple, as long as you know the drive is gone. But what if you don't know? What if nobody notices the little red light on the front of the box, or complains about a slow response? Eventually another drive fails, and you lose the entire server. I hope your backups are good.

You could visually check your servers every day for little red lights. While you're there, you might as well log on and check how much disk space is left, and if any odd things show up in the event log. I don't know about you, but I'd rather be getting stuff done. Computers are much better at eternal vigilance than we are, so let's get a good monitoring package and let our servers watch our servers. When the first symptoms of a problem begin to show, the software can yell for help.

System Monitoring Software

There are a number of products out there for the purpose of server monitoring, and they all perform the same basic function: Keep track of your servers' vital signs and notify you when something starts to go awry.

Around that formula, however, there are many variations. Some products focus more on network services, making sure that websites are still up, print servers are still serving, and servers are still alive on the LAN. Other products let you peek into the internal functions of your machines, to monitor disk, CPU, and memory attributes.

Alerts tend to be logged to a central location for later review, but can usually be sent by email and pager. Some products offer additional methods of delivery, like cell phone text messages, instant messages, and maybe even carrier pigeons.

Vendors have mostly developed and targeted their products specifically for one of three types of customer: The large enterprise; midsize business; or small business market. Packages intended for larger businesses are more complex and customizable, and are generally priced accordingly.

Category One - Big Bucks For Big Business

Major vendors have built their business around immense "Network Management Systems" costing hundreds of thousands of dollars in software, hardware, consulting, and staff. HP's OpenView, CA's Unicenter, and Concord's eHealth are major products in this arena.

These products are all highly customizable and feature a common framework into which different modules are installed. Each module performs some specific task. One module might monitor database performance, for example, while another measures traffic on network segments.

Generally, a company purchases the modules it needs, and then expends an enormous amount of effort configuring those modules to match the environment to be monitored. Intelligent and well-planned deployment of one of these Network Management Systems can significantly enhance the efficiency of a big company's IT operation.

Poorly planned execution, however, can turn into a quagmire. I was caught up in one such project several years ago that generated only resumes. More and more people were assigned to work on getting the monitoring software working and stable, leaving less and less to actually react to the problems being reported.

Category Two - Midsize Shops, Middling Budgets

The next tier down offers a good mix of products in the four-figure range covering most aspects of systems monitoring. Products such as Ipswitch's WhatsUp Gold and Quest's Big Brother provide a good mix of features for midsize networks.

These products offer log monitoring, bandwidth utilization, and automatic event handling in addition to basic metrics like CPU utilization and disk space. More specialized products are also available. NetIq's Webtrends product, for example, focuses exclusively on Web server monitoring, and its detailed and flexible reports make it a popular tool for both IT staff and Marketing folks.