Archive for October, 2008

Coming Soon: Spreadsheet Views

Tuesday, October 21st, 2008

If you’ve been reading our blog regularly you’ll have noticed that v5.3 is going to include a lot of new functionality.  We’re happy today to announce another great feature that’s coming your way in this release.

In the Views section of the tree for devices and event monitors you’ll find a new item:

Device Spreadsheet View

Device Spreadsheet View

The spreadsheet views show your devices and event monitors in a display that looks just like an Excel spreadsheet.  You can select rows, columns or areas of cells and then copy them to the clipboard in multiple formats.  You can sort by any column just by right-clicking on the column title.

Device Spreadsheet View

Device Spreadsheet View

And of course you can export the data to Excel or any other spreadsheet program so you can integrate it with your other data and systems.  In our internal testing we’ve already found that the spreadsheet views make it really easy to get big picture of our monitoring configurations, especially those with a hundred network devices or more.

Posted by Don

Coming Soon: System Health Event Monitor

Thursday, October 16th, 2008

In almost all monitoring environments the five most import things to keep track of tend to be:

  • CPU Usage
  • Drive Space Usage
  • Memory Usage
  • Bandwidth
  • Ping Response Times

Of course, Tembria Server Monitor has supported all of these since its initial release way back in 2003.  All you had to do was create one event monitor for each type and then set your thresholds to define how you want to be monitored.

Today we’re happy to announce a new event monitor that brings all of these items into one event monitor that is super easy to setup.  Say “hello” to the System Health Event Monitor:

System Health Event Monitor Settings

System Health Event Monitor Settings

Along with this event monitor we’re including a brand new dashboard element that collects the data generated by the event monitor and shows it in an intuitive interface.

If gives you an at-a-glance view of systems that might be under stress, with those that might need the most attention listed first.

This example here shows drives ranked by their available space:

System Health Dashboard

System Health Dashboard

You can also create System Health dashboards elements for CPU, drive space, bandwidth and memory usage and then arrange them using our integrated dashboard editor.

The new event monitor and dashboard element will be included in v5.3 which is due to ship at the end of the month.

Posted by Don

Tip: Managing Event History Size

Monday, October 6th, 2008

The integrated event history is one of the most fundamental parts of Tembria Server Monitor.  When an event monitor runs it generates one event history record for each network device that it checked.  The event history record tells you whether or not the check was successful and the event text tells you what was checked and what the result was.

By default, Tembria Server Monitor keeps a maximum of 50,000 event history records.  When the limit is exceeded the newest event history record pushes out the oldest one.  You can change this number by going to the “Event History Settings” in the “Settings and Preferences” section of the tree.

Event History Settings

Event History Settings

So what’s a good value?  On average hardware, say a modern desktop PC with 2G of RAM and a single core processor, we recommend keeping the value to 200,000 records or less.  On a higher end system like a quad-core server class machine you can bring it up to perhaps 750,000 records or more.  The exact reasonable value will depend on the machine in question and on your monitoring configuration too.

That said, we’ve noticed though that many users aren’t aware of the various options we offer to manage the creation of event history records.

Let’s say you’re using our Ping event monitor to watch 100 machines every 30 seconds.  You’re going to be generating a tremendous number of event history records with the default settings.  The important question is:  Are you ever going to want to look at all of these records?  It’s unlikely that you could, even if you wanted to.  You’re likely only interested in the ones that failed either due to a machine being unreachable or a response time threshold being exceeded.

In this case our recommendation is to set your event monitor to only write to the event history on failure or, if you prefer, on state change.  With the “on failure only” option you’ll only get event history records for failed checks.  With “state change only” you’ll get one record when a device check fails and one when it comes back online.  And here’s the part that’s not obvious:  You’re still collecting graph data for all the other checks.

Event Monitor Notification Styles

Event Monitor Notification Styles

So you get the best of both worlds.  Your event history will be  uncluttered and easy to browse because it only contains records for important events, plus you get the long term trending data in the graphs.

Posted by Don