cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

BT Contribution Split Calculation and Reset

jason_ostroski
Dynatrace Pro
Dynatrace Pro

When utilizing the Contribution splitting restriction, how often is this updated and/or reset?

I have a very particular BT I am working with to try to show the transactions that are the biggest response time contributors for an application. I would like to be able to show the top average worst transactions of the day/week. Utilizing the top X filter will not work as there are too many transactions for the DT server to be able to analyze in a reasonable amount of time and reports usually come back blank when I try to run this data (full analysis for 24 hours of data is estimated to take 4 days). I am splitting the BT by eventSource because this is only way identify the hundreds of unique transactions for this app.

My idea is that I can utilize the contribution restriction to only show the worst transactions for the app. However, I would like to know how often this is updated and/or reset. From the documentation:

When you configure a splitting restriction, the splitting values of an application are determined based on an initial set of transactions. This set is one of the following, depending on the load:

  • All transactions in an observation period of 1 minute
  • The first 500 transactions that are collected, if fewer than 500 transactions occur within 1 minute

With a sufficient load, there can be more than 500 transactions in the set. The final splitting values are calculated by taking the contribution of every splitting value of a transaction into account. Splitting values with a small contribution are mapped to a remainder splitting value.

Does this mean that this is never updated, it only uses the initial configuration and I would have to reset the splitting to have it re-analyze the transactions?

Thanks,

Jason

PS: If you have a better workaround for my scenario, please let me know.

2 REPLIES 2

andreas_grabner
Dynatrace Guru
Dynatrace Guru

Hi Jason

You interpreted the doc correctly. Once we have our 500 splittings we dont update it any longer. This is not ideal for a moving data set. For what you want to achieve I think the best is to use our Real Time Business Transaction Feed to e.g: Elastics Search or Splunk. You can stream out every single split value and then use the power of these BI solutions to analyze Top X for Timeframe Y.

Here is the link to Real Time BT Data Feed

Thanks Andreas, I wanted to see if there was a solution within the tool, but already have had conversations with the Splunk team at my client to test the real time BT feed for this reason.

Thank you!