Memory growth in fleet adapter and traffic schedule (#664)

[Charly_Wu]

Hi,

The above chart is with this implemented to prevent task log growth.

In my development environment, I used heaptrack and at the same time run script to collect VmRSS from /proc/$PID/status and the charts are as shown below.

heaptrack:

The dips are when cache is cleared.

VmRSS:

[grey]

Ah nice, does your heaptrack data include the Top-Down call stacks that led to the memory allocation? That would be the important information that we need to track down what’s causing memory to build up.