Our third monitoring dashboard entry!

We’ve had another great entry for our dashboard competition, this one from Jonathan Allen. This submission contains some more conceptual designs, looking at possible approaches to representing high-level status information on a dashboard. There are three parts to the entry:

Aero View of Monitored Entities

This view would not be dissimilar to the Aero view in Windows Vista – i.e. a way of easily moving between the detail displays for the different servers. Jonathan explains:

“I think this offers a good surface area for the server you are interested in so plenty of detail can be on display. The view should switch to standard windowed format once a server is selected so that investigation and correction can take place”.

AeroStyle

Galileo Thermometer View of Alerts

A key point about this view is that the monitoring application is providing an alert-centric view of current problems rather than a server-centric view. I.e. it is prioritizing alerts by their severity, regardless of which machine they are on. The alerts would rise to the top as they became more severe – the operator would then need to prioritize these items above the lower, less severe alerts.

GalileoThermometer

Spacemonger View of Alert Priority

Again, this is a view prioritising the severity of various alerts over the machines on which they’re found. As Jonathan notes:

“In Response v2 I would suggest the size and colour of the server’s rectangle would indicate its need of attention – large + red = urgent, smaller + green = no rush. Within that rectangle you can then prioritise by type of alert – deadlocks are red and larger than long running sql that is green … All rectangles allow drill-down so that the clicked rectangle becomes that full UI area and more details appear.”

spacemongerview

We’re intrigued by these entries; as the name of this website suggests, we’re interested in the future of monitoring, and in particular the problem of how to prioritise different problems across your server space. Many dashboards provide a very familiar interface, for example showing a range of standard charts often featuring the red/amber/green coloring system. Do you prefer a familiar layout so that you don’t need to learn a new user interface or do you think there is a missed opportunity here for something new and innovative?

A key issue raised by these designs is that the last two provide an alert-centric view rather than a server-centric view. Is this what you prefer? Is there a danger that a severe alert on a development box is prioritized over a medium level alert on a production server – even if the problem with the latter should be dealt with first? Is there a need to prioritize different servers (i.e. “Any alert on this production server is more important than even a high-level alert on my dev boxes”).

What do you think? Let us know!

This entry was posted in Design A Dashboard. Bookmark the permalink. Post a comment or leave a trackback: Trackback URL.

5 Comments

  1. Tom.Randle
    Posted October 6, 2009 at 10:55 am | Permalink

    Drawing the user’s attention to the most serious problem can be a difficult problem.

    Raising the most critical issue to the top, like in thermometer, or devoting more screen space are both quite nice ways of achieving this.

    The next issue is defining what the most serious or important problem are!

  2. Posted October 6, 2009 at 5:07 pm | Permalink

    Red and Green? NOOOOOOOOOOOOOOO! As one of the 1 in 4 men who are colour blind, I often find myself squinting at monitoring displays, trying to work out if everything’s fine or everything’s really bad because someone else has decided to use a traffic-light system…

    • Adam.Walker
      Posted October 7, 2009 at 12:55 pm | Permalink

      Thanks Gareth. So have you found any displays that work better for you? Would something like the Galileo Thermometer View, where issues rise to the top according to severity, work better? We’d be interested to hear your thoughts.

    • jonathan allen
      Posted October 9, 2009 at 1:33 pm | Permalink

      Hi Gareth – a fair point well made.
      I too have a mild issue with Red/Green but I used them in the suggestion as it seems to be a universal Red=Bad, Green=Good format (traffic lights etc).

      I would hope if the interface is using lots of colours that the palette will be customisable for differing needs. Grey scale, thickness/style of border, pattern of fill and so forth can be alternatives or in addition to colour. Screen location and size(screen area) are indicators that can show importance and are not dependent on any colour scheme (traffic lights – stop if the one on top is on, go if the one at the bottom is on …).
      Mouse hover tooltips would also give instant advise on how the user should react – any text that begins “WARNING – …” would need an urgent click whereas a text starting “Information – …” can be sidelined and cleared at a more sedate pace.

      Thanks for your comments

  3. Merrill Aldrich
    Posted October 9, 2009 at 9:37 pm | Permalink

    These are cool ideas – goes to the fact that the dashboard is 90% an information design problem and only 10% a technical/development/programming problem :-)

Post a Comment

Required fields are marked *

Add an Image

*
*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>