Bugs in Infrastructure as Code

Abstract: Infrastructure as code (IaC) scripts are used to automate the maintenance and configuration of software development and deployment infrastructure. IaC scripts can be complex in nature, containing hundreds of lines of code, leading to defects that can be difficult to debug, and lead to wide-scale system discrepancies such as service outages at scale. Use of IaC scripts is getting increasingly popular, yet the nature of defects that occur in these scripts have not been systematically categorized. A systematic categorization of defects can inform practitioners about process improvement opportunities to mitigate defects in IaC scripts. The goal of this paper is to help software practitioners improve their development process of infrastructure as code (IaC) scripts by categorizing the defect categories in IaC scripts based upon a qualitative analysis of commit messages and issue report descriptions. We mine open source version control systems collected from four organizations namely, Mirantis, Mozilla, Openstack, and Wikimedia Commons to conduct our research study. We use 1021, 3074, 7808, and 972 commits that map to 165, 580, 1383, and 296 IaC scripts, respectively, collected from Mirantis, Mozilla, Openstack, and Wikimedia Commons. With 89 raters we apply the defect type attribute of the orthogonal defect classification (ODC) methodology to categorize the defects. We also review prior literature that have used ODC to categorize defects, and compare the defect category distribution of IaC scripts with 26 non-IaC software systems. Respectively, for Mirantis, Mozilla, Openstack, and Wikimedia Commons, we observe (i) 49.3%, 36.5%, 57.6%, and 62.7% of the IaC defects to contain syntax and configuration-related defects; (ii) syntax and configuration-related defects are more prevalent amongst IaC scripts compared to that of previously-studied non-IaC software.