* Add container_cpu_cfs_* metrics (CPU throttling due to limits)
* Add container_memory_swap metric
* Ensure minimum kernel version for thin_ls

Diff: https://github.com/google/cadvisor/compare/c6c06d4...2ed7198