Diagnose Root Cause

Smart Assisted Triage workflows designed to help the user to quickly diagnise root cause through problem correlation, propogation and metric analysis.
Contributing Members
  • Sruthi Iyer: Sr. UX Designer

The future direction is for ATC to be smart and present the potential root cause of issues to the user before they occur, so the issues can be resolved proactively.

Presently ATC aims to assist the user in identifying the root cause by and triaging alerts in the system along with an extensive palette of details and allows for the user to the sift through the various pieces of evidence available and deduce the root cause of issues before they cause more damage.

The workflow is meant to help the user with accomplishing the following tasks:

  • How do I know which problem I should focus on next?

  • How do I find the cause of this problem and go through a streamlined support workflow to see the problem fixed?

On examining the shortcomings of this current workflow, we realized users are unable to accomplish their task with ease because of these concerns:

  • I cannot correlate problems with each other because I cannot see how the pieces of evidence associate or developed over time.

  • There is too much information on the Notebook view, and I’m not sure how to go about discovering the root cause.


Once the users are aware of the problem they need to address, there is massive cognitive overload and confusion when too much information is exposed with no “help” or nudging from ATC toward what the user should focus on or explore further.

There are many compelling and powerful (feature-filled) views available within the Analysis Notebook but usability has suffered because users are having to “learn” how to use these different UIs to their advantage. The usability issues are exacerbated by the fact that when the Notebook is opened for an Experience card that contains too many components, the load times for each of these UIs are different and can be too long most of the time (especially for the topology). This results in users learning a few UIs that they are comfortable viewing (that load fast) and not bothering with the others. A big part of the feature set goes overlooked or accessed at only when there is no other way to get to the required information.

A significant improvement in usability can be achieved by segmenting the parts of the notebook to help drive the user through a deliberate set of steps of exploration-investigation tasks while continually maintaining the environment and time context.

The bulk of the workflow which is to: prioritize -> Explore correlations -> Investigate causesremains across these three screens. The environment and time contexts are fixed at the top across them. A fork in the workflow can be created to dig deeper into any of these views by an advanced user but the workflow progresses primarily through the content exposed in these three views. This way, the task can be accomplished by navigating the product through insights from the content rather than learning the UI.