Change from load to linux pressure metrics for better per resource usage monitoring

we are using the some avg300 pressure indicator from the linux kernel to determine the availability of resources instead of the old load value which was a collection of multiple load sources and didn´t accurately show usage.