Search code examples
javanewrelic

Understanding Top 5 requests by percent of wall clock time graph for newrelic

Just wondering how to understand top 5 requests graph for newrelic. How does wall clock time get defined? I had some theories but seeing more than 100% is throwing me off completely. I am using newrelic java agent.

enter image description here

UPDATE

As per @lauradiane answer: Concurrent requests can make percentage go above 100% Say I have 5 transactions for the REST endpoint I am looking at and each takes 12 seconds to finish. Lets assume that I am looking at 60 seconds window.

So for concurrent execution of these 5 transactions:

% of wall clock time = (5x12)/60 x 100 = 100%

If they execute serially

% of wall clock time = 12/60 x 100 = 20%
Solution

  • Per New Relic doc:

    We look at the wall-clock time (the amount of time as recorded by the clock) for all the transactions, and we sum that across all of the transactions. Because the server can execute requests in parallel, you end up with percentages over 100. A percentage of 100 would would mean that the execution time across those 5 transactions is equal to the time expended when recording wall clock time.