Hi Heiko, there is a glitch with -1 versus 0 values in mod_cluster web console, it looks like
-1 values are being clipped to 0.
|Node Timeout: 0||node-timeout=-1||-1|
|Worker Timeout: 0||worker-timeout=-1||-1|
|Flush Wait: 0||flush-wait=-1||-1|
Hmm, as long as smax=-1 (Soft maximum idle connection count = no idle connections), ttl=-1 (Time to live (in seconds) for idle connections above smax) seems legit.