Across all nodes, and all job types, ~1/100 jobs that are run will hang on the executor. 

The Job will complete, if the job is clicked on the log shows completion alongside the green checkmark. However when looking at the Node the job ran on, one of the executors still shows its being used and keeps an active track of the time. (in the picture attached the job has been "running" for 3 days, but actually finished 1.5 hours after launching)

This was has only been seen on Jenkins 2.426.3

These hung executors then need to be cleaned manually, 1 by 1, via a groovy script.

2 pictures attached that show what I just described.