On the Crash Stats page, there are different stats broken down by device type, OS version, time of day, country, etc. Each metric has a blue and gray line underneath. The blue line represents the percentage of total crashes belonging to the attribute of focus, and the gray line represents the percentage of total sessions belonging to that same attribute. Large differences show where crashes are over-indexed for a particular attribute, helping your team to better prioritize and discover crashes that are impacting subsets of users. Look for cases where the blue bars are significantly longer than the gray bars.
For example, looking at Device Distribution, specifically Lenovo TAB 10 devices, we see a long blue line and a short gray line. Upon hovering over Lenovo, a pop-up appears.
The popup message should be interpreted as follows for the device type; Lenovo TAB 10. Note, the data displayed are tied to the time period selected. Update the timeframe on the right-hand side of the page to see data for a specific period.
Count: 165, the number of crashes that occurred on that device; Lenovo TAB 10
Affected: 8.45%, the percentage of total crashes that occurred on that device; Lenovo TAB 10
Overall: 0.28%, the percentage of total sessions within the selected time period that belonged to the device, Lenovo TAB 10
Difference: +8.17%, the difference between the % of Affected - % of Overall of Sessions. Large differences signal higher correlations, and in this case that Lenovo TAB 10 is more likely related to crashes.
*If both the blue and gray bars are close to being even, that means that for that attribute (in the example below, United States), crashes and sessions occur with similar frequency.
