Retry handler JBOSS server on TCP port remote hos

Elastic Stack Logstash
1

HI,

i have one big issue, indeed, i have one ELK on 1 standalone host (A)....

Then i have 2 jboss (wildfly) cluster different hosts (B,C) and for each one one logstash handler which send on remote host logstash listener (A) on port 5001 for example.

i have to start first the listener(A), and then the web servers (B,C), it is a big issue, because for example if the listener (A) crash due to OOM (out of memory) or other bug, then after restarting it (A) log will not be caught until we restart all webservers (B,C), it is not acceptable, have you a solution to configure on logstah handler (B,C) ? or logstash listener (A)?

thank you i have to implement it but i can't propose this solution due to this issue...

2

As Logstash here acts as a server and does not initiate the connection, you probably need to change the behaviour of the clients so that they attempt to reconnect if the connection is lost.

3

in fact in wildfly i have update the logextmanager librairie (2.0.4.Final instead 2.0.0.Final) in system module and now there is a connection retrying... so i will now use logstash-forwarder (shipper) because with a handler we loose some start and stop logs...

if topic can be closed the problem is fixed, it is a server lost connection

thank you !!