Disk queue length metric
Atlas exposes a few hardware metrics including Util% (presumably obtained from iostat or similar). However this metric is not very significant in the age of SSDs - a volume could be 100% utilized and still have spare capacity. A more useful storage-level metric is Average queue length - this is easier to interpret (high queue length = storage contention).
Would it be possible to add this metric to Atlas monitoring?
11
votes
Sylvain
shared this idea