Currently, at the default logging level, we are not logging entries on when Logstash attempts to re-establish connections (eg. due to DNS for downstream servers not available). For example, after these messages when the DNS issue is resolved:

2017-09-07T02:50:43-07:00 WARN DNS lookup failure "<hostname>": lookup <hostname> on <ip>: read udp <ip>:51430-><ip>: i/o timeout
2017-09-07T02:50:43-07:00 ERR Connect failed with: lookup <hostname> on <ip>: read udp <ip>:51430-><ip>: i/o timeout

Having this information logger at the default level can help with troubleshooting.