UPDATE 

I have found out that:

• for total_duration/val = 1, the corresponding transaction average trt will be equal to max trt and min trt, suggesting that the quotient (total_duration/val) is the number of data points that were averaged to get average trt..
• the quotient can go beyond 2 (I had some go as high as 9 in my recent load tests)

It would seem that the quotient is the number of vusers executing a corresponding transaction..

For example, for a particular transaction, let the individual trt (in seconds) of four Vusers be 2, 1, 3, and 5. This gives a total of 11. 

Their average trt would be 2.75. And quotient 11/2.75 = 4. Ergo, quotient = number of Vusers executing a transaction.

I'll leave this here for someone to correct me if I am wrong about this.

The data is aggregated, there is a value per interval (5 seconds interval). 

Say LRC captured the following three response times of the same transaction within an interval: TRT1=1sec, TRT2=2sec and TRT3=3sec. In this example, the duration of the three transactions is summed up to 6 seconds, 1+2+3=6. Thus LRC will report in the CSV "total_duration" = 6" 

The "value" for this interval is the average of the three transactions measured during the interval, 6/3=2 seconds. Thus LRC will report in the CSV "value" = 2". 

I hope this is clearer, we will enhance the LRC documentation as needed. 

Thanks!  

Much clearer, thanks! 

Although in case you'd still check this thread.. is there documentation on how LRC computes the values displayed in the Report? We are trying to code the computation (using the csv) ourselves and include it in our pipeline..