Virtual Machine Explorer

 

Changes and Issues for any virtual machine in your datacenter can be tracked in time and correlated with performance metrics for effective root cause analysis. Our vision for this feature was to enable admins to dig deep into problems uncovered in their monitoring tools by exploring the current and historical condition of the virtual machine and its related objects such as Hosts, Clusters, Datastores to detect patterns. 

The mockup below was an early version of the VM exploration mode to communicate our vision. It depicts the ability to search and open a detailed view for any type of virtual object. Multiple views open up as tabs. Every object in our platform will be a link to their respective detailed view. The time navigation allows the user to see issues, changes and configuration for any object within the time range selected.

 
 
CloudPhysics_VM_Explorer.png
 
 
 

The two data grids for Changes and Issues have expandable rows to reveal rich information about the Events. As part of these designs I also created the underlying data model for Issues. Issues have severity and certain information associated with them such as the Date, Title, Severity, Description, Impacted Objects, Recommendation and Remediation. The designs show how this information would be presented in the context of a single object such as a virtual machine.

 
 

Iterations

 

Since releasing the first version of the Virtual Machine Explorer, we've iterated regularly to incorporate additional features, style enhancements and usability improvements. Some of our original assumptions on the data also changed which meant adapting the design accordingly. For instance, we discovered that unlike "Change" events, there's not enough value in presenting a historical view of "Issues". The work required in preserving this data is significant and costly.

This was a major change that implied I needed to edit the design to align with the new assumptions. In my original design, I presented the Issues and Changes tables side by side and the data presented in these tables was based on the time range selected by the user. For the new assumptions, time tracking was only applicable to changes and the performance charts. "Configuration" and "Issues" would be based on the latest snapshot.

Our exploration view went from being entirely anchored in time to primarily representing the current snapshot with a time exploration capability comprising a chart view of performance metrics annotated with configuration changes.

I adjusted the layout to make this clear.

 
VM_explorer_updated_Oct30_2015.png
 

In addition to major layout changes, I continued to push my developers to follow design guidelines around alignment, spacing, typography etc. We tested the screen for various browser resolutions and sizing to apply scrolling as necessary and prioritize the 4 main containers in a way that all the necessary information was accessible. We set min-heights on the chart so that resizing the browser did not make it unusable. 

 
I captured a screenshot of the version in dev and marked it up for changes.

I captured a screenshot of the version in dev and marked it up for changes.

 
 

The final round of changes included meticulously designing the tabbing feature. Tabbing in this view is akin to tabs in browsers as opposed to fixed tabs commonly used for navigation.

 
 

latest version in production