We successfully implemented a blue-green deployment strategy for our SAP S/4HANA quote management module running on SAP BTP. The business requirement was clear: zero downtime during releases, even during peak sales hours.
Our approach involved maintaining two identical production environments (blue and green) with automated health checks and instant rollback capability. We integrated this with our CI/CD pipeline to enable seamless deployments. The blue environment serves live traffic while green gets the new release, then we switch routing after validation.
Key benefits we’ve achieved: deployment time reduced from 4-hour maintenance windows to 8-minute switchovers, zero customer impact during 12 releases over 6 months, and automated rollback within 90 seconds when issues are detected. The solution uses SAP BTP Cloud Foundry routing with custom health monitoring scripts that validate quote creation, pricing calculations, and approval workflows before switching traffic.
This transformed our release velocity and eliminated the stress of coordinated downtime windows with sales teams.
How do you manage stateful sessions during the switch? If a sales rep is actively creating a quote when you flip from blue to green, do they experience any disruption?
Our rollback is primarily routing-based, switching traffic back to the blue environment within 90 seconds. However, we implement careful data versioning strategies. Quote records include a schema_version field that indicates which application version created them. Both blue and green environments can read all versions but only write in their designated format. If rollback occurs, the old environment continues processing without data loss. We also maintain detailed transaction logs during the switchover window for audit purposes. Critical validation includes testing quote creation, modification, approval workflows, and pricing engine calculations before declaring the green environment healthy.
What’s your automated rollback mechanism? Does it just switch routing back, or do you handle data consistency issues? In quote management, if quotes were created in the new version with different data structures, rolling back could cause problems.