The ones and zeros are barely dry from the first PerfStack release, and already we have slew of new features and improvements lined up that we're eager to get feedback on.
When PerfStack was initially released, a surprising few key metrics were noticeably absent. Chief among those, was the surprising lack of Syslog data. In this release of PerfStack we rectified this injustice, bringing Syslog into the PerfStack fold. For any node sending Syslog data to Orion you will find a new 'Syslog' metric tile under the 'Status, Events, Alerts' category. Simply drag that tile to the chart area, and voila! Syslog data charted over time broken down by severity level. As with all metric tiles within PerfStack this metric tile is dynamic and will only appear Nodes which have been configured to send Syslog data to the Orion server. If this tile does not appear for nodes you believe should, verify that Orion has received Syslog data from the device using the web based syslog viewer within the Orion web interface.
Hovering your mouse over the charted area will show the total number of syslog messages received at the top of the legend for the time shown below the vertical marker, as well as a breakdown of each severity type contributing to that total beneath it.
Not to be outdone, the inclusion of SNMP Trap data has also been added to this release of PerfStack. Much like Syslog, SNMP Trap data is broken out by severity and is available as a dynamic metric tile for any node which is configured to send SNMP Trap data to Orion. The SNMP Trap metric tile can be found under the same 'Status, Events, Alerts' category as Syslog, when a node is selected from within the metric palette. If this tile does not appear for nodes you have configured to send SNMP Trap data to Orion, verify that Orion is receiving those SNMP Traps from the device using the web based SNMP Trap viewer, accessible from within the Orion web interface.
When hovering your mouse over the charted area you may notice your mouse cursor has learned a new trick, changing to crosshairs. Holding down your mouse button while dragging across the charted area allows you to lasso a specific time period of interest, such as a sudden spike in resource usage. The selected area is then the focal point of your project, while the rest of the chart area surrounding this time period becomes visually de-emphasized through color desaturation. while the colors remain bright and vibrant within the selected area. This easily allows you to focus your attention on the area selected area without visual distraction from the surrounding area. This selection process occurs across all charts within the PerfStack, making it easier to visually correlate what else occurred during the same time period.
Once a time period has been selected, new options appear next to the selected area. Clicking the [+] icon highlighted above, zooms into the selected time range where you can view high fidelity detailed data collected during that time period. Similarly, clicking the [-] icon allows you to zoom further out to spot trends, or return to your previous perspective after having zoomed in. Any time after having made a selection, clicking the [X] icon will cancel the selection and return focus to the entire viewable chart area.
Great! So now you know how many Syslog messages and SNMP Traps were received from any of your devices, along with their respective severity. You can even zoom into a specific time frame and cross correlate this data against other metrics from the same or different entities monitored by Orion. That's some super powerful stuff! But you know what would make this data even more powerful?
What if you could actually view the full details of those Syslog or SNMP-Traps from within PerfStack itself? That would assuredly accelerate the troubleshooting process and aid in reducing time-to-resolution; both of which are key tenets of PerfStack. Well that's exactly what we've done!
To get started, select a time period in the chart area, the same as described above to zoom in or out of the chart. Next, click the top paper & magnifying glass icon from the options displayed to the right of the selected area. This action will cause the Data Explorer tab to be shown in the left pane, where the Metric Palette normally appears. Switching between the Data Explorer and the Metric Palette is simply a matter of clicking on the appropriate tab.
Within the Data Explorer, all Syslog or SNMP Trap data is listed in the chronological order it was received. Each line represents a single message, beginning with its severity and ending with the date and time the message was received. In between is a brief preview of the message body. To view the full message text, simply expand the row.
If your devices are sending a lot of Syslog or SNMP Traps to Orion, it may be difficult to sift through all the noise and focus on what's truly important, even if you select the tiniest window of time. Since this just so happens to be another key tenet of PerfStack, we added filtering and search to the Data Explorer. This allows you to do things, such as show only Warning or Critical messages that were received during the selected time period, while filtering out the excessive noise generated by Informational and Debug messages.
And BOOM! Just like that, the problem is found. If SolarWinds was a leading US based office supply company, right now you'd probably feel compelled to blurt out 'That was easy". The magic doesn't end there either. This same functionality that's available for both Syslog and SNMP Traps can also be used to view the details of Orion Events. So if during the beta you aren't afforded the opportunity to configure devices to send Syslog or SNMP Traps to your Orion server, you can still give this new feature a whirl by adding Orion Events to your PerfStack and following the instructions outlined above.
Visualization of alerts in the premiere release of PerfStack displayed only the total aggregate of all alerts against a given entity, as well as how long all alerts had been active against that object. While useful, you were unable to determine what those alerts were which triggered, or how long each of those alerts remained active. A lot of that important detail was obscured through aggregation, making this particular area of PerfStack ripe for improvement.
In this release of PerfStack we preserved the same total alert aggregate that was available in the first release, but then extended the chart to show the names of the alerts which triggered along with their individual durations. No longer are you left wondering what alert was triggered against the object, or fumbling through other areas of the Orion web interface to track down that information. You'll also know at a glance when the alert triggered and for how long, in a manner which allows you to visually recognize patterns of reoccurrence and correlate specific individual alerts against other metrics collected by Orion.
Occasionally it's necessary to share your PerfStack findings with others who may not have access to Orion, or archive those findings in a ticketing system for historical purposes beyond your defined retention period. For some, this information needs to be imported into other systems for change back, show back, billing, forensic analysis, or correlation with other non-Orion tools in the environment. Previously, your only recourse was to export data for each metric individually via the Custom Chart View, or write your own series of custom queries against the SQL database backend to obtain the raw values behind these charts. In this PerfStack release however, those days are behind us.
After adding all metrics of interest to your PerfStack project, simply click the new 'Export' button located in the top menu. This will export your project's content to a Microsoft® Excel friendly comma separated value (CSV) formatted file,which is then downloaded to your local machine via your browser.
Double click the file to open in Excel, or upload the file to Google™ Docs to open in Sheets. Inside you will find all the raw data which made up the charts in PerfStack. Each series of the same chart is represented as its own set of columns, complete with the entity and metric name, date and time the data was collected, and its value. If your PerfStack project contained multiple charts, the values for each chart are grouped together in the CSV file the same as they were visually laid out within PerfStack itself. Simply put, the detailed raw data for the second PerfStack chart is directly beneath the end of the series for the first chart data contained within the CSV file.
The logical layout of the CSV file makes it easy to visually recreate similar charts to what was seen in PerfStack, using the native graphing tools included in Excel or Sheets. The format of the raw data is also well suited for import into 3rd party reporting solutions like Domo or eazyBI. With the data contained in logical layout and an open format, the possibilities are limitless.
Not everyone who used PerfStack initially realized that whatever they created could be easily shared with others, simply by copying the URL and sending it to them. Since this is one of PerfStack's most powerful capabilities, we felt it needed to be promoted to the top menu, alongside other important functions like saving and loading a PerfStack.
Just click this 'Share' button and the dynamic PerfStack URL is automatically copied to your clipboard and ready to be pasted into an email, instant message, helpdesk ticket, etc. and shared with others.
Believe it or not, we still have more in store for PerfStack, but in the meantime we'd love to hear your thoughts on the improvements we've made above. Do you have a favorite new feature? Did we miss the mark? Please provide your feedback below and let us know what you think.