Symptom Radar Chart is a machine-learning powered feature designed to help explain the symptoms behind at-risk drives. It characterizes each drive’s health along 5 symptom axes: S.M.A.R.T., Temperature, Drive-Detected Issues, Self-Test, and Host-Detected Issues. These axes are then displayed on a radar chart.
Each axis can be thought of as summarizing a group of symptoms that were collected in a particular way. For example, the axis called S.M.A.R.T. was generated by taking values from S.M.A.R.T. attributes and feeding it into a machine learning model that outputs one value summarizing the intensity of several S.M.A.R.T. Attributes. The symptoms that feed into the S.M.A.R.T. axis do partially overlap with the symptoms that feed into the other axes. But the key here is that the symptoms that feed into each axis is collected via a distinct mechanism.
A lower axis score indicates poorer health along the corresponding health axis. And the lowest score approximately corresponds to the decile survival rate over the next year. For example, if the lowest axis score is 2, the drive only has a 20% chance of remaining with the user over the next year according to historical data.
Clicking on an axis will highlight the corresponding symptoms at the drive level, making it easy to isolate active symptoms.
To see this feature:
1) log in to DA Portal
2) navigate to an individual drive
3) click on the Threshold Based Alerts tab
4) click on any axis to highlight the symptoms that belong to the axis on the page