Department of Computer Science: MSc Thesis Presentations

Mitigating Configuration Drift in Infrastructure-as-Code Systems

Author: Otso Pohjola

Supervisor: Lachlan Gunn

Abstract: Terraform is widely utilized in cloud ecosystems, which evolve and expand quickly and are prone to configuration drift, leading to increased complexity and uncertainness. Some solutions to this exist already, but they are either platform dependent, or utilize third party tools to achieve this. Terraform's native toolset has been expanded lately, offering new options for combatting drift, but these require manual effort.

 We developed a solution that expands Terraform's own change management and that automates Terraforms native drift mitigation operations, by generating a Terraform import resources of identified unmanaged objects. The import resources allow embedding present unmanaged objects into Terraform state and configurations.