If I understand it, the percentage is recalculated every leak check. So
it only reflects the instant io flow, instead of historical statistics?
But I think for system admin purpose, it's good to know a longer range
io activity character. Or do you think management tool should sample it?
Yes I
On Tue, 07/23 18:29, Benoît Canet wrote:
This metrics show how many percent of the time I/Os are put on hold by the
throttling algorithm.
This metric could be used by system administrators or cloud stack developpers
to decide when the throttling settings must be changed.
Signed-off-by:
This metrics show how many percent of the time I/Os are put on hold by the
throttling algorithm.
This metric could be used by system administrators or cloud stack developpers
to decide when the throttling settings must be changed.
Signed-off-by: Benoit Canet ben...@irqsave.net
---
block.c