Abstract

As container orchestration environments, like Kubernetes, get more popular due to their simpler deployment, problems arise. With this, certain requirements for those orchestration environments become essential, one of which is the reliability of the software services it contains. This thesis presents recommendations that help to achieve a reliable software service, whereas I am focusing on stateful services as those are more complicated. Furthermore, I choose a definition of software reliability, considering failure scenarios, and provide concepts to prevent them. Moreover, I put the concepts to the test, which gives an insight into whether these recommendations increase the reliability of software services and which obstacles arose during the implementation of our recommendations.