Hi Stijn,
I have occasionally been seeing similar behavior when profiling one of
our Solr 3.6.1 servers using the similar AppDynamics product. Did you
ever hunt down what was causing this for you or get more info? (I
haven't been able to rule out truncated or filtered call-graphs that
don't show some very-large number of sub-threshold calls... but that
seems implausible)
Always looking to learn more,
Aaron
---original message---
Hi,
I'm working with a recent NightlyBuild of Solr and I'm doing some serious
ZooKeeper testing.
I've NewRelic monitoring enabled on my solr machines.
When I look at the distribution of the Response-time I notice
'SolrDispatchFilter.doFilter()' is taking up 90% of the time.
The other 10% is used by SolrSeacher and the QueryComponent.
+ Can anyone explain me why SolrDispatchFilter is consuming so much time?
++ Can I do something to lower this number?
( After all SolrDispatchFilter must Dispatch each time to the standard
searcher. )
Stijn Vanhoorelbeke
