ISPPlanet NMS Series - Ipswitch WhatsUp Gold 5.0

Service Monitoring

WhatsUp Gold can also keep an eye on Internet services running on networked devices. Monitor listening ports for SMTP, POP3, FTP, HTTP, NNTP, IMAP, DNS, Time, Telnet, Gopher and UDP Echo by configuring the device Services panel (right). Most ports can be scanned and monitors enabled with ICMP discovery. Otherwise (e.g., after SmartScan discovery), you must scan ports for each device individually to complete monitor configuration. Thereafter, service availability is checked during each poll interval, if the device is reachable.

WhatsUp Gold signals service failure with one color—by default, purple. Service failure is propagated onto the next-level map as yellow. But this limits ability to assess severity—yellow could indicate a cranky device harmlessly missing a few polls, or a mission-critical server with applications on the fritz. Services are not included in device availability statistics and are not easily factored into dependencies. We spotted one way to overcome this: poll a device with TCP and one service. If this service goes down, the "device" goes unreachable. The ability to specify service severity (e.g., consider device X unavailable if either HTTP or FTP are down) would make WhatsUp Gold a stronger application-level monitor. If your map has many submaps, or you primarily plan to monitor services, you'll prefer the Services window (left).

If you want to monitor a service that is not in the basic list, configure a custom service. A handful of "custom" services are included: an HTTP content monitor that checks a URL, and RADIUS and SSL monitors. We easily configured a new FTP content monitor, using the Custom Service panel (below, left) to specify port, timeout, and expect/response strings. A rule editor can be used to create complex regular expressions. Examples assist with rule design, but you must understand the protocol being added. Log and debug viewers let you see how your monitor is behaving, invaluable during rule development (below, right).

WhatsUp Gold also provides a COM (Component Object Model) "plug in" API to add non-TCP custom monitor types. If you're a Win32 programmer who knows COM, you can write your own "plug ins" to monitor odd devices or create more advanced application monitors. Two freely-available "plug ins" illustrate this concept: an NT/2000 Services monitor and an SNMP Threshold Monitor. WhatsUp Gold must be momentarily shutdown during plug-in installation.

We used the NT/2000 Services monitor to track Informix and RFC1006 status on a remote NT server. Simply supply the name of the server, and WhatsUp Gold displays available NT services. WhatsUp Gold must of course run under a user with administrative permission to access remote NT services.

We used the SNMP Threshold Monitor (below) to check ifOperStatus on our routers. This monitor significantly extends WhatsUp Gold's SNMP capabilities. Use this monitor to trigger service-level events, color changes, and alerts by thresholding the absolute or relative value for any MIB object. ISPs can use this monitor to be notified when link, CPU, or disk utilization become excessive, or when availability falls below contracted levels. The only serious limitation spotted here is the one noted for SNMP grapher: objects must be qualified by instance, requiring separate "custom monitor types" for each table row/column.