ISPPlanet NMS Series - Ipswitch WhatsUp Gold 5.0

Trap Monitoring

WhatsUp Gold can monitor SNMP agents by listening for TRAPs (right). Incoming traps are signaled subtly by highlighting the device's label rather than changing icon color. Trap variable bind lists can be viewed from the device's Log panel or from the overall Log Viewer. As we'll see, traps can be used to selectively trigger alerts. But that's the full extent of WhatsUp Gold trap handling. This product does not factor trap severity into device status. It cannot ignore traps based on community string. Trap viewers cannot be searched, and traps cannot be selectively cleared. If you want a more sophisticated trap handler, write one yourself (invoke another program upon trap receipt) or use another product.

Alerts and Notifications

According to Ethernetix, a Chicago high-speed Internet access provider, WhatsUp Gold notifications make it possible to offer better customer service.When a customer's T1 was accidentally disconnected, WhatsUp Gold paged Ethernetix founder Richard Dib. Dib called the customer and opened a telco trouble ticket, using the circuit ID sent to his pager. "I was able to alert the customer that his Internet was down even before he knew," said Dib. WhatsUp Gold supports a wide variety of notification types (left), including audible, pop-up window, pager-beeper, voice message, and email notifications.

Devices are configured, individually or as a group, to raise alerts based on thresholds and triggers (right). Thresholds determine how many failed polls must occur before alerts are logged. After reaching the threshold, triggers are used to invoke notifications. For example, ignore one failed poll, log if at least two polls fail, send email when a fourth poll fails, and call someone if the device remains unreachable for 10 poll cycles. Notifications can also be triggered by SNMP traps—any trap or specific trap(s), filtered by id.

Notifications can also be grouped together, in series or in an escalating sequence. For example, page the day shift from 9 to 5 or the night shift from 5 to 9. Try to page first. If the page fails, try email. If the email message can't be sent, raise an audible alert. Customizable detail can be included with nearly any notification. WhatsUp Gold provides a lot of flexibility here. It also does a good job of logging alerts, both failed and successful, so that you can later see who was or wasn't notified, and when.

WhatsUp Gold provides basic support for automating alert response. It can invoke any application on the WhatsUp Gold system, supplying alert detail (e.g., date/time, device name, IP, failed service name, trap variables) as parameters. The NT/2000 Services monitor can restart a failed NT service. Unfortunately, we found NT service restarts did not raise alerts or generate log messages—the offending service was simply restarted. This could allow a festering problem to go unnoticed.

WhatsUp Gold's notification support is very strong, but we were perplexed by inconsistent notification behavior encountered when monitoring multiple services. For example, when we monitored SNMP and HTTP on host X, notifications were only triggered if both services were unreachable. But when we monitored SNMP and FTP on the same host, notifications were triggered when either service went unreachable. Ipswitch support tells us the latter behavior is expected and is investigating our results.