put a rate-limit on truncate attempts

[amir-arad]

in TruncatingFileWriter when failing to truncate, the next log line will re-attempt the truncation. in high volumes this can add load to the disk and might escalate the problem.

I think we should put a rate-limit on the truncation attempts (a default of 1 second between attempts?) and log the failures as they may provide valuable insight to the system's state. preferably to the other log writers as well.