Automatically stop tagged EC2 instances when AWS billing crosses a set threshold.
Add a spending circuit breaker as a safety net against runaway AWS costs.
Limit blast radius by requiring instances to be explicitly tagged as stoppable.
Enforce the stop permission at the IAM level so untagged instances are protected.
| bradleygamimarques/lizard | anas727189/devsecops-pipeline | edwinjdevops/damolak-challenge | |
|---|---|---|---|
| Stars | 0 | 0 | 0 |
| Language | HCL | HCL | HCL |
| Setup difficulty | hard | — | hard |
| Complexity | 4/5 | — | 3/5 |
| Audience | ops devops | ops devops | ops devops |
Figures from each repo's GitHub metadata at analysis time.
Requires enabling billing alerts in the payer account and deploying in us-east-1, since billing metrics only exist there.
Lizard is an infrastructure-as-code project for AWS that acts as a spending circuit breaker. It watches your AWS billing through CloudWatch Alarms, and when your estimated charges cross a threshold you set, it automatically stops EC2 instances that you have tagged as expendable, so a billing spike does not keep growing unattended. The idea came from a real AWS billing bug on July 17, 2026 that showed some accounts estimated charges in the billions of dollars. Lizard was built so people could have a safety net running in the background: if spend crosses a line, the resources marked as stoppable get shut down. The readme is upfront about what this tool can and cannot do. It cannot figure out which specific instance is causing a cost spike, because a billing alarm only reports that spending went over a limit, not which resource caused it. Instead, you tag the instances you are willing to sacrifice with a specific tag, and when the alarm fires, every instance carrying that tag gets stopped, not just the one responsible for the extra cost. Instances are stopped rather than deleted, and the permission to stop them is enforced through AWS identity and access management rules, not just through the automation script, so an untagged instance cannot be stopped even if something in the automation goes wrong. Setting it up requires knowing that AWS billing metrics are only available in the us-east-1 region, and that the payer account must have billing alerts enabled before Lizard can see spending data at all. The project uses Terraform for its infrastructure code and Yarn plus a tool called lefthook for contributor tooling, and it is currently a small, early project with a single author.
A Terraform project that auto-stops tagged AWS EC2 instances when your billing alarm shows spending has crossed a threshold.
Mainly HCL. The stack also includes Terraform, AWS, CloudWatch.
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly ops devops.
This repo across BitVibe Labs
Verify against the repo before relying on details.