Logging issue with Amazon Linux 2

[lynchem]

Since updating one of our deployments to Amazon Linux 2 we have some issues with the log output. The biggest issue is that if you dump an exception or callstack to the console that feeds into cloudwatch logs as multiple lines rather than as previously as a single line. Another irritation is that you end up with a prefix on every line (although this is easier to remove).

Feb 6 20:16:28 ip-172-31-21-148 web: 

See the screenshot

We did some digging and the only reference to it we can find it is this serverfault post

We pipe our logs into elastic and this throws off all the filtering and dependent alarms.

Anyone seen this or know of a workaround?