Lambda trace metrics don't match Cloudwatch

[j_butzow]

Hello - 

We have some Java lambda functions that are instrumented for tracing with Dynatrace OneAgent.  However, the metrics from Cloudwatch do not match what Dynatrace is reporting about Lambda function invocations.

In particular, the 'code execution duration' metric for the lambdas is much higher than the duration of any distributed trace (e.g., duration is reporting 6 seconds, but the distributed trace is ~18ms).  Is this a common problem, or is anyone aware of why there is such a gap between these values?

Thanks!

[ChadTurner]

@j_butzow are you able to provide screen captures of the Dynatrace Data and the CW Data? I'm curious as to why you are seeing this and if the some type of average/aggregation might be playing a role in the altering of the values form one tool to the other.