Add error_output to eventlogging service and make eventbus write EventErrors to log file
This is just extra insurance against failed events. In case we have a Kafka hiccup,
this failed events log file will allow us to re-produce failed events manually
after the hiccup has passed.