Handle EC2 Spot Interruption Notices Gracefully
Poll the instance metadata for an interruption notice and drain gracefully so a spot instance is not killed mid-request.
Task
Configure curl -s http://169.254.169.254/latest/meta-data/spot/instance-action so the cloud resource enforces the control automatically without manual console steps.
When to use this
An audit finds that curl -s http://169.254.169.254/latest/meta-data/spot/instance-action was configured manually and drifted, leaving a resource exposed for two weeks.
Prerequisites
- Terraform or IaC tooling installed
- Cloud provider credentials with target permissions
Steps
1. Configure curl -s http://169.254.169.254/latest/meta-data/spot/instance-action in the project configuration file with the recommended defaults.
curl -s http://169.254.169.254/latest/meta-data/spot/instance-action
Expected output: The configuration is accepted and curl -s http://169.254.169.254/latest/meta-data/spot/instance-action runs without errors on the first invocation.
2. Run the check locally to verify curl -s http://169.254.169.254/latest/meta-data/spot/instance-action produces the expected result before pushing.
pnpm test
Expected output: Local run matches CI behavior; no surprises after push.
3. Add the gate to the CI pipeline so curl -s http://169.254.169.254/latest/meta-data/spot/instance-action blocks merge on failure.
- name: Run add ec2 spot interrupt notice
run: pnpm test
Expected output: Pull requests with failing curl -s http://169.254.169.254/latest/meta-data/spot/instance-action are blocked from merging.
Troubleshooting
Problem
The check passes locally but fails in CI
Solution
Pin dependency versions and the runtime to match CI exactly; reproduce with the same Docker image or act.
Problem
Configuration is rejected with a schema error
Solution
Check the tool version matches the config schema; older versions may not support newer fields.
Limitations
- The gate only runs on events that trigger the workflow; locally pushed commits bypass it.
- Configuration drift is possible if someone edits settings outside version control.
Codify the control
Start by measuring the problem. Poll the instance metadata for an interruption notice and drain gracefully so a spot instance is not killed mid-request. The most common mistake is adding the gate without first confirming it catches the issue you care about. Run curl -s http://169.254.169.254/latest/meta-data/spot/instance-action against the current codebase and inspect the output before wiring it into CI.
curl -s http://169.254.169.254/latest/meta-data/spot/instance-actionDetect drift
Once the gate is in place, make failures actionable. A failing curl -s http://169.254.169.254/latest/meta-data/spot/instance-action should point to the exact file and line, not a generic error. Pin the tool version in the lockfile so a upstream release does not silently change behavior. Review the gate output in every pull request just like a test result.