You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Queries handled by the packet cache don't seem to end up in the ringbuffers. This makes dynamic query counting, e.g. via exceedServFails(), useless for the usual repeating packet floods by misbehaving clients on similar FQDNs.
While it certainly has some performance impact I wonder if it would make sense to adjust the ring buffers so that they can also be used for proper counting of queries hitting the caches, potentially in a configurable way. Or look into a different solution for being able to handle excessive clients/queries without causing excessive performance impact.
If kept as is I think the current behavior might be missing some documentation at ringbuffers / dynamic blocks.
(As discussed)
The text was updated successfully, but these errors were encountered:
Short description
Queries handled by the packet cache don't seem to end up in the ringbuffers. This makes dynamic query counting, e.g. via exceedServFails(), useless for the usual repeating packet floods by misbehaving clients on similar FQDNs.
While it certainly has some performance impact I wonder if it would make sense to adjust the ring buffers so that they can also be used for proper counting of queries hitting the caches, potentially in a configurable way. Or look into a different solution for being able to handle excessive clients/queries without causing excessive performance impact.
If kept as is I think the current behavior might be missing some documentation at ringbuffers / dynamic blocks.
(As discussed)
The text was updated successfully, but these errors were encountered: