How do you approach testing and monitoring solutions built with Infrastructure-as-Code tools like Terraform?
Implement monitor_terraform_runs(runs, failure_limit) to process Terraform run records. Each record contains a workspace and either "success" or "failed"; return workspace names when they first reach failure_limit consecutive failures, preserving alert order. A successful run resets that workspace's streak, and each workspace should appear at most once.
def monitor_terraform_runs(runs, failure_limit):