feat: More accurate endtime when log hits max size

[lcottercertinia]

Is your feature request related to a problem?

if an event has a start (METHOD_STARTED) but not end event, the end time of an event becomes the very last timestamp when the log reaches *********** MAXIMUM DEBUG LOG SIZE REACHED ***********
This makes it look like the method too a long time e.g 30 - 40 seconds.

However in reality we have not idea how long it took since we do not know exactly where it ends in the *********** MAXIMUM DEBUG LOG SIZE REACHED *********** gap.

Describe the solution you'd like

Instead set the the end time of these to be the last time we have before the gap e.g the end time of the last child.
The FATAL_ERROR that comes after *********** MAXIMUM DEBUG LOG SIZE REACHED *********** should be moved to under the root since we do not know where it occurred.
In addition the *********** MAXIMUM DEBUG LOG SIZE REACHED *********** seems to be sometimes placed in the wrong place of a log so we should move it to the correct place e.g after the last good event and before the fatal error.