Use Case

As a Data Explorer developer I want deployments to automatically rollback when automated checks fail So that bad deployments never reach production and user experience remains stable

User Experience

• Users experience fewer incidents from bad deployments
• Faster incident response - automatic instead of manual intervention
• Increased confidence in deployment process

Risks/Unknowns

• Current bake period comes after deploy to each environment - may not integrate simply with rollback
• Need to align with SRE vision and workflow template capabilities
• May require rethinking the deployment pipeline structure

Acceptance Criteria

Implementation Requirements

• Enable auto-rollback using new batch-deploy workflow template
• Integrate with existing bake period that occurs after deployment to each environment
• May not be as simple as passing an option to the workflow
• Coordinate with SRE team to ensure alignment with their vision

Testing Requirements

• Try to deploy a failing image?

References

• https://jira.mongodb.org/browse/SRE-493