There's more...

A metric that is often computed along with MTBF is MDT. MDT is the average amount of downtime (repair time) that is required in order to fix a failed machine or system. To compute MDT, do the following:

1. In Power BI Desktop, create the following measure:

MDT (Hours) = 
    VAR __Table = 
        ADDCOLUMNS(
            'R02_Table',
            "__Hours",
            DATEDIFF(
                'R02_Table'[RepairStarted], 
                'R02_Table'[RepairCompleted],
                SECOND
            )/3600
        )
RETURN
    DIVIDE(SUMX(__Table,[__Hours]), COUNTROWS(__Table), 0)

2. On the same Report page we created earlier, create a card visualization and place the MDT (Hours) measure into the Fields area of the visualization.
3. On the same Report page, place the MDT (Hours) measure into the Values area of both table visualizations.

Now we can see that certain machines and causes of failure take longer to repair than others.

Finally, to demonstrate the predictive value of MTBF, do the following:

1. In Power BI Desktop, create the following measure:

Next Expected Repair = 
  VAR __LastRepair = MAX('R02_Table'[RepairCompleted])
RETURN
  __LastRepair + [MTBF (Hours)]/24

2. On the same Report page we created earlier, add Next Expected Repair to the Values area of the table visualization that includes the MachineName column.

Now, we can see by MachineName when the next expected failure event is likely to occur.