Self-Healing CI/CD
D.A.M.I's flagship feature — automatically detect, diagnose, and fix CI/CD pipeline failures.
How It Works
Supported Scenarios
| Scenario | Detection | Auto-Fix |
|---|---|---|
| Dependency conflicts | ✅ | ✅ Version resolution |
| Test failures | ✅ | ✅ Suggest code fixes |
| Docker build errors | ✅ | ✅ Dockerfile patches |
| Infrastructure drift | ✅ | ✅ Terraform plan + apply |
| Security vulnerabilities | ✅ | ✅ Dependency upgrades |
| Configuration errors | ✅ | ✅ Config correction |
| Resource exhaustion | ✅ | ✅ Scaling recommendations |
Approval Gates
D.A.M.I never makes dangerous changes without your approval:
- READ operations (logs, metrics) → Execute immediately
- WRITE operations (create PR, update config) → Confirmation card
- DESTRUCTIVE operations (delete, rollback) → Explicit approval required
Setting Up Self-Healing
- Connect your CI/CD tool (Jenkins or GitHub Actions)
- Configure webhooks to send failure notifications to D.A.M.I
- Set approval preferences in Settings → Governance
- That's it! D.A.M.I will automatically respond to failures
Example: Jenkins Pipeline Failure
When Jenkins sends a build failure webhook:
{
"name": "deploy-service",
"build": {
"status": "FAILURE",
"number": 142,
"full_url": "https://jenkins.example.com/job/deploy-service/142"
}
}D.A.M.I will:
- Fetch the build logs via MCP
- Analyze the failure using Gemini AI
- Identify the root cause
- Suggest or apply a fix
- Create an incident record with full timeline