Fluentd failing with Segmentation Fault

Issue

• Fluentd is trying to connect to our external log store, but is constantly reconnecting:

2021-08-11 18:12:18 -0400 [warn]: dead connection found: ocp-logging.example.com, reconnecting...
2021-08-11 18:12:23 -0400 [warn]: recovered connection to dead node: ocp-logging.example.com
2021-08-11 18:12:23 -0400 [warn]: disconnected host="ocp-logging.example.com" port=9624

• Fluentd logs show Segmentation Fault:

2021-08-11 17:20:24 -0400 [warn]: disconnected by Error error_class=Errno::ECONNRESET error=#<Errno::ECONNRESET: Connection reset by peer> host="ocp-logging.example.com" port=9624
/usr/share/ruby/openssl/buffering.rb:319: [BUG] Segmentation fault
ruby 2.0.0p648 (2015-12-16) [x86_64-linux]

• Buffers are getting full; when the buffer on the pod reaches 256 MB it is no longer able to forward any logs.