how many listen directives is too many for haproxy configuration on a single node? What are the most important resources to monitor?
is 50 times the
listen qwerty213
bind *:9124
mode tcp
option tcplog
balance leastconn
server zxcv 10.20.0.13:31527 check inter 3000
[...]
too many? Or 100? (each one has several backend servers too)
Is there any limit after which haproxy performance deteriorates due to having a great number of open local ports which are connected to backend servers?