We would like to have good logging and monitoring/alerting from day 1 on our kubernetes based services. For this to happen we should figure out a way to do this without impending the developer much. There are a number of approaches on this one, from enforcing using specific preconfigured frameworks like service-runner, to creating sidekick containers that run alongside the "main" container (but in the same pod) and e.g. consume the main containers stdout for logging and poll it for statistics gathering. There is also the issue of encryption. In an ideal world, our pods would be able to terminate TLS as well as have all outoing connections encrypted. That may or may not be possible directly on the main container and may or may not make sense to be implemented via a sidekick container.
This task is about discussing our approaches, picking the best one and implementing it.