The Enqueue and Dequeue count metrics reported by the MQMonitor Agent drop to zero between the times that our agent polls for data. This gives a false impression that these counts are zero which they are not, it’s just that the agent is not collecting data during those intervals. We are requesting that the data point jump from one collected value to the next value when it is collected, rather than drop to zero.
The normalization performed in the code is impacting this.
A desktop build removing normalization and adding new metrics was created (Enqueue & Dequeue Counts /Per Hour/Per 6 Hours/Per Day). If no metric, the count will not fault to zero unless actually is zero. Ask Support to see if DE332914 applies for your situation. This will be in APM 10.7.1
Some additional fixes were added in DE363041.
Problem 1: Thread timing problem:
APM was polling MQ for metrics without considering the previous poll duration. Due to this, APM was not correctly polling for metrics at exact delay configured (say for 60 seconds delay configured, APM was polling at 60 seconds + previous poll duration, which is incorrect)
Problem 2: When metric value drops to lower value, the new value is not reported immediately. It is taking 15 seconds delay.
Problem 3: Metric counts are not reported accurately when metric value fluctuates