Challenges

This project was originally intended to run with kubernetes. Due to limited server resources we changed course to use docker compose instead. Kubernetes was more likely to consume storage by writing to logs that aren't directly associated with kubernetes and are not removed when kubernetes data is cleared. This logging issue was first discovered after the switch to a docker compose based configuration. Due to the late integration phase of the project, there wasn't enough time to react adequately to this requirements change. This led to a time-consuming administration job because instances of services first need to be stopped, their images deleted and then new pulled instead of just pulling and replacing a running service. Furthermore logs needed to be limited to a small size which made debugging an interactive job because missed error logs were overwriten not even a minute later.