I am working to understand the impact of different workloads on the load weight and WALT (ravg) in the linux scheduler. Specifically, I am working on the Android Kernel.
Currently, I am logging all of this data in an internal struct from inside the kernel. For the particular workload, I write the aforementioned data into the struct every time the pick_next_task is called.As one would expect, the size of the logged data increases significantly within a short time period. Hence, I flush the data by writing it to file after the size of a log increases to a certain value.
I wanted to ask if there is another, better, way to log this time series data?If not, then what would be a good way to stop the size of the log from exploding? Is there an alternative to writing the data to file? Would using debugfs be a better alternative?
Any help would be greatly appreciated.Thank you!