The JVM max heap is set to 512M. The K8s deployment memory is set to 1G for its limit and 1G for its request size. The Grafana (Promethus data) is accurate and reflects the 512 value roughly as expected. Dynatrace on the other hand shows over 688M for heap & 423M used. These are vastly different numbers than what Grafana is showing.
I would like to understand what data DT is showing and why are we seeing such a difference. Eden and Tenured graphs look similar between the two sources.