Investigate memory leak

[x-xo]

We have experienced the node getting restarted in ECS which seems to be rooted in a memory leak, or similar.

At the time we hit the ~100% the node got restarted and the drops thereafter have been caused by (intentional) restarts.
However, the chart suggests that the node is eating up memory until it exhausts the available resources.

We should check if we can spot the source of this.