We have multiple JMS consumers (IBM MQ) , listeners read messages from IBM MQ and does the business logic which includes calling multiple java methods and database calls. Dynatrace is automatically detecting the IBM MQ Listener and also JMS consumer and is correctly showing the purepaths but the actual throughput shown on dynatrace for those processgroups or listener is 5-6 times more compared to actual number of requests processed by listener .
For ex listener is processing ~ 6K requests per hour whereas dynatrace is showing 30K per hour which is causing discrepancy and inaccuracy of the data.
Ticket here: SUP-46026
Is there any additional settings we need to do inorder for dynatrace to correctly report the stats.
Dynatrace Managed 1.190
3 node cluster with Active Gate
Any chance you count just a subset of the messages? Did you look at the PurePaths?
Regarding message listeners - I've seen implementations where also "dummy" messages were processed. They, of course, resulted into PurePaths and into higher throughput in Dynatrace. The customer was only counting the non-dummy messages.
issue is resolved, problem was due to server adaptive load reduction happening frequenntly on the Dynatrace managed cluster which results in multiplication of purepaths, dynatrace support was able to quickly identify the issue and adjusted the cluster threshold as well as updated cluster to latest version 1.92 which have improvements to ALR Algorithm
This does not sound like the ALR is kicking in. In case ALR is activated, the number of captured PurePaths (traces) will be lower, however, the number of requests must match. ALR does not lower the throughput numbers for a service.