Disk run out of space while running a build , We do use Disk Usage plugin , but it checks for disk usage every hour also mark the slave offline without any notification which is not appropriate
The plugin checks for disk available before starting a job and fail the build is the disk is below the Threshold ( You define the Threshold in a global parameter) ,
Also there is a Global Configuration Enable Disk Recycler if it's checked and the disk space is low for a job , then it will delete everything in the current job workspace directory,