*5.2. Planning*

The independent variables in the experiment are the number of simultaneous pod replication requests made to the service, overloading the Nginx application server, and emulating Kubernetes' autoscaling so that the service continues to be available under high workloads. The dependent variables were: CPU utilization, memory consumption, disk utilization, and average response time to requests.

Following the GQM method, the following research questions were designed to broadly cover the scope of this work:

• Q1?: Have indicators of software aging been found?


In order to answer Q1 and Q2, the following metrics were evaluated: CPU utilization, memory consumption, and disk utilization. In order to answer Q3, we conducted a comparative analysis of the results obtained from metrics that have been used to answer Q1 and Q2.
