Hi R Marc,
you may try the snippet below to collect some
LB_FAILED
information.
when LB_FAILED {
log local0.debug "Client: [IP::client_addr]:[TCP::client_port], Pool: [LB::server pool], Online Pool Members: [active_members [LB::server pool]], Member: [LB::server addr], Member-Status: [LB::status]"
}
But I guess it would never trigger on the node specific connection limits, unless every single pool member has exhausted its limit (in this case the entire pool will be down). If just a single member has reached its limit, then the member would be just skipped during LB selection and causing no errors...
You could also try to monitor the
PERSIST_DOWN
event, in the case you're using a session table based persistence method. This event should trigger for those clients, that have an active persistence entry for the node that has just reached its connection limits (or is somehow marked as down because of other reasons).
when PERSIST_DOWN {
log local0.debug "Client: [IP::client_addr]:[TCP::client_port], Pool: [LB::server pool], Online Pool Members: [active_members [LB::server pool]], Member: [LB::server addr], Member-Status: [LB::status]"
}
Cheers, Kai