Hi,
After quite a bit of debugging, I've found that our Logstash performance cripples moving from 6.2.4 to 6.3.0 (and have tested up to 6.5.4, all being identically poor)
We're using RabbitMQ as input and elasticsearch as output, with a lot of logic in filters.
With 6.2.2 we're seeing about 6-7.000 messages processed per second, moving to 6.3.0 or beyond, these numbers fall to 50-100 messages processed per second.
I've looked at the Logstash release notes and can't find a smoking gun other than some generic work in the multi pipeline area.
Anyone know what can cause such great degradation?
Cheers, Bjorn