Shared Terraform Module Library for a Data Platform
One reviewed security baseline for every team's data lake resources
Overview
At a telecommunications company, each team wrote its own Terraform for Lambda, S3 and Glue. Every version needed its own security review, and a fix made by one team never reached the others. I built a shared module library that defines the security baseline once, so a hardening change is reviewed once and reaches every team with a version update.
Problems
Each team defined the same types of resources separately, so each one had to be reviewed separately
A misconfiguration or outdated pattern in one team's code was never caught by a fix made somewhere else
Security and compliance reviews were repeated for every implementation instead of being done once against a standard
With no central owner, known fixes stayed with the team that happened to apply them
Goals
Replace duplicated resource definitions with one centrally maintained standard
Make sure a fix applied once reaches every team that uses the module
Cut the review workload from one review per team to a single review of the shared module
Make secure configuration the default result of using the module
What Was Done
Replaced independently written Lambda, S3 and Glue definitions with one maintained standard
Set up a model where one hardening fix updates every team's infrastructure
Reduced the review surface for these resource types from one implementation per team to one shared module
Results and Impact
Security review of these resources now covers one shared standard instead of every team's copy
A vulnerability or misconfiguration fixed in the module is fixed everywhere it is used
Least-privilege IAM and encryption at rest come with the module rather than depending on each team
The organisation has a single auditable source of truth for these resource patterns
Stack
