ccwells

Sparklines (UMP List View Display Type 'Spark')

Discussion created by ccwells on May 12, 2011

Within the Nimsoft Unified Management Portal, the List View applet provides a tabular, elastic view of your QoS data (Example: VMware ->https://skitch.com/ccwells/r6h34/list-viewer).

 

Tabular - rows and columns. Typically rows show host or target (i.e. the entity you've monitored) and the columns represent the metrics gathered for that host (CPU, Disk, etc.)

 

Elastic - as hosts come and go (particularly useful for data from the Nimsoft vmware probe) the List View is updated automatically.

 

Among the many display options for each column is something called 'Spark' - "intense, simple word-like graphs."

 

I was curious about the best use of this display type and found the following link to a blog post by the foremost authority on information presentation, Edward Tufte:

 

http://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=0001OR&topic_id=1

 

The key thing I noticed is the necessary context for the Spark - simply noting the metric doesn't quite get it:

 

https://skitch.com/ccwells/r6h56/spark-no-context

 

A simple change to add the most recent value makes an improvement:

 

https://skitch.com/ccwells/r6hh4/spark-with-context

 

Hopefully you find this helpful.

 

~/charlie


Outcomes