Master master relationship getting "too many connections" error after extended outage

I have a master master database setup with an onprem and cloud hosted MariaDB server. I can stop one database or the other for testing purposes and start it back up and everything seems fine. The server that remains up will continue to serve connections when the other is down. However, we had an extended power outage recently causing the connection between the two to break for around two hours. After around an hour our cloud hosted server started throwing "too many connection" errors. Is it possible that the cloud server kept trying to connect to the onprem master/slave and it resulted in hitting a limit on the number of connections? All the retry and timeout values are default. When the connection was back up, restarting the service in the cloud resulted in the sync catching up and everything working. However, I'm looking for a way to keep my cloud hosted server serving connections even if the onpremise server is down for an extended period.