In my case, I was using a custom index, that index had no route and such would trigger the exception case code. The exception case bug still exists and still masks the real issue, however I was able to work around this by testing the connectivity with other tools such as nc -vzw1 myindex.example.org 443
and retrying when the network was up.