Adaptive Sampling Latency Spike in Dynatrace OneAgent 1.239 Under High Concurrency

Goal: Identify the concurrent request load at which Dynatrace OneAgent 1.239’s adaptive sampling mechanism begins to add measurable request latency, given that the effect is driven by thread‑contention detection and varies with host CPU capacity.

Constraints: The adaptive sampling threshold is not exposed in the UI, its default value differs across environments, and the latency increase only appears under concurrent requests; disabling adaptive sampling removes the spike but may reduce metric fidelity. Because the exact concurrency level depends on the number of available cores and the current workload, a universal threshold cannot be stated without further measurement.

What concurrent request count triggers adaptive sampling on a specific CPU core count? How does adjusting the adaptiveSamplingMinInterval setting shift the latency onset? Are there differences in concurrency sensitivity between OneAgent 1.239 and later releases?