I run the nagios nrpe server on a Lightsail Debian instance. This needs a custom port opening, TCP 5666.
Last year, without warning communications between the nrpe client on a remote host and the nrpe server running on the Lightsail instance started to fail.
Further investigation revealed that the remote client and the Lightsail instance were still talking on the custom port, but both sides of the connection were now reporting that the SSL handshake was failing.
When I disabled SSL on the nrpe server running on the Lightsail instance everything worked fine.
No updates had been made to either the remote client machine or the Lightsail instance that the nrpe server is running on. The same remote client machine could still successfully connect using SSL to multiple other nrpe servers not hosted on the AWS network and local SSL connections to the nrpe server on Lightsail also worked fine.
In summary the only failed connections to the nrpe server on Lightsail were from remote clients using SSL.
The next day when I checked, everything had mysteriously started working again with successful remote SSL connections to the nrpe server on the Lightsail instance.
Today, exactly the same thing started happening again.
Is this some wierdness on the AWS network? Has anyone else experienced this?
Edited by: EIC-BTC on May 25, 2021 9:01 AM