[dnsdist] increase in drops or changed semantics?
cm at appliedprivacy.net
Mon Oct 7 20:47:00 UTC 2019
>> Has the way how drops are counted changed recently?
> No, I don't think we changed anything in that area recently. The drops
> counter is the number of queries sent to a backend and for which we did
> not get a response fast enough (or at all).
> You should be able to use the inspection features of dnsdist to figure
> out what is causing these drops, like grepq("3000ms") or topSlow().
The output of topSlow() including the 'Rest' line covers only about 1%
of the count shown in the 'Drops' counter so it is not related or is
topSlow() only about recent-ish slow queries?
(there is no column description in the output, I assume the 3rd column
is number of requests?)
More information about the dnsdist