I am currently investigating some memory issues with our Node.js container instances and noticed that during memory peaks, a vast amount of data is transferred between AWS instances and the Atlas primary replica member.
Is it possible to analyze from which collections in Atlas the traffic is originating?
The only thing I know is that we are reading from primary only and that the traffic is incoming from the container perspective, i.e. it is flowing out of Atlas.
Is there any metric in Atlas which helps to analyze the traffic flow?
It would help us much since the traffic flow correlates with the memory issues we try to debug.
Best,
Martin