Jun 28, 2014
Hard to compare fairly; not only do we need comparative experience with the leading platforms it's important to know what YOUR needs and environment are. Barring that, we can provide our experience....
No experience with either Oracle or Catchpoint, I do have a bit of experience with HP Openview (long ago) and mrtg, and a lot more current experience with Solar Winds Orion, now known as NPM.. With these 3 you sort of get what you pay for, but the curve isn't linear. Openview is certainly the most expensive. MRTG really should be teamed with something else to get more in depth with systems, but it seems to be easy to get network statistics from it. I have not seen anyone use it to pull system process details, but perhaps someone with extensive experience here can chime in.
Which brings us to NPM. It has matured a lot in the past 4 years. Originally just upgrading was a major project and fraught with errors. These days the upgrades are smooth, the product has more power and it's less effort to maintain. Like Openview most of the power is in modules purchased separately, so it is important to decide up front what you need in terms of monitoring and how many nodes. So, for example the base product will monitor a set number of devices, drives and network ports/adapters. Additional modules are available for application monitoring, SAN monitoring, endpoints and more. Distributed/larger networks will need additional monitor engines, another factor in the total cost.
Like most monitors there is an auto discovery mode that shortens the configuration time. Be aware that even so there is substantial effort required to configure any network monitor. The default views are useful, but creating custom maps and page views really makes it easy to spot problems at a glance. The map creation tool is easy to use and is intuitive, but mapping your network so it makes sense to your staff is necessarily a creative task. You may want to set up an initial simple map, then review after a few months to make improvements.
Creating alerts is likewise fairly simple, but the key is knowing what is important as well as what constitutes a critical level that requires notification. For example, our T1 over IP link shows high utilization by design; including it in a high utilization alert is pointless. Likewise printers may need monitoring but we don't want status alerts at night when they might be turned off. Fortunately alerts can be set for time and day of the week to prevent unnecessary alerts. Advanced alerts are created in a similar way with a separate tool, but rely on conditionals to provide much more power. They also have quite a range of actions beyond just sending an email or text alert. How much you choose to do will depend on your needs and the time available. Just remember to lean on other users who share considerable knowledge including custom pollers for less common hardware and more.
Disclaimer: I do NOT work for Solar Winds or any other vendor and YMMV. Corrections for errors and omissions gratefully accepted.
