ISPPlanet NMS Series - Castle Rock SNMPc

Trend Reporting

Basic on-demand reports can be generated from any device's right-click menu (right). These tabular or graphic reports, which can be printed or saved to file, document average response time and availability for each device and monitored service. An important caveat here: don't confuse SNMPc port scan stats with true application-level service monitoring and "content verification."

In addition, custom trend reports can be configured by adding nodes to a Trend Selection Tree (left). Choose from more than a dozen built-in report types like Interface Health and Web Response Time. Define the device(s), SNMP object(s), and interval used to collect trend report raw data. In theory, an "Instances" panel lets you be selective about saved data — for example, record only non-zero counters. While we could open this panel, we were unable to really use it, and the manual indicates that Instances are an Enterprise Edition feature.

In the Workgroup Edition, Trend Reports can be exported to a comma or tab-separated file on a regular schedule. A program can even be executed whenever data is exported (e.g., launch Excel, convert data to another format). In the Enterprise Edition, more comprehensive reporting features are available, including scheduled ODBC export, printed, and web (HTML) reports, refined by filters and layout parameters.

Trend reports can be displayed on demand by both versions (right). A calendar is used to select the period (days, months) to be included in the report. Options determine how data is aggregated: a single merged graph for all devices, a graph per device, or a graph per variable. As with any other SNMPc graph, on-demand reports can be printed or saved to a file.

SNMPc can trigger threshold alarms based on observed trends (left). When this feature is enabled, SNMPc creates a "baseline" by averaging polled objects over a "learning period". Once a baseline and threshold have been established, SNMPc generates a warning if the object's value exceeds this threshold. Thresholds can be automatically adjusted up or down over time, and warning events can be limited (per minutes, per object). Rules can of course be defined to automate threshold event handling. The attached example illustrates an auto-generated FTP threshold, generating one warning per hour when more than 10 percent of FTP polls fail for a given device.