[Pdns-users] causes of over-capacity-drops

[David]

On 2018-08-15 3:11 AM, mendisobal via Pdns-users wrote:
> Hi, I use pdns-recursor (4.1.2-1pdns.xenial, Ubuntu 16.04.5 LTS), it 
> serves 700 qps. Several times a day the over-capacity-drops counter 
> increases to 12 qps at this time the concurrent-queries counter does not 
> increase. Please tell me the reasons for this phenomenon and how to 
> eliminate it.

Concurrent-queries is just a point in time reference and it can change 
wildly between when you grab your other metrics.

With such low QPS I'd just do a tcpdump and reference when it happens. 
You should see a thousand or more uncached queries being sent from 
something which would be what ends up causing the over capacity drops.

concurrent-queries > threads*mtasker count = over-capacity-drops

The typical reasons for this are (a) lots of clients that have their own 
cache - and they refresh some common name all at the same time when the 
TTL expires or (b) abusive traffic.

> 
> 
> _______________________________________________
> Pdns-users mailing list
> Pdns-users at mailman.powerdns.com
> https://mailman.powerdns.com/mailman/listinfo/pdns-users
>