Open ierceg opened 7 years ago
Considering that we are capturing engine container's output for analytics, we should ensure that logs of engines that weren't captured by lazy (e.g. a defect in log capturing code) are captured on the next start.
Instead of streaming which is prone to issues, simply use a log scraper like fluentd.
Considering that we are capturing engine container's output for analytics, we should ensure that logs of engines that weren't captured by lazy (e.g. a defect in log capturing code) are captured on the next start.