Databases

Sanitas achieves database modernization with a true DevOps operational model

October 11, 2023

Matthias Kupczak

Data Management Specialist

Florian Graf

Google Cloud Consulting

Sanitas, a regulated Swiss-based healthcare insurance company with over 800,000 customers, is decidedly not a startup. However, the spirit of innovation remains core to its mission of customer-focused service.

In 2020, Sanitas embarked on one of its biggest projects yet by migrating their customer-facing technology infrastructure to the cloud. After years of working with on-premises servers, the company wanted to help its developers innovate faster, based on cloud principles.

One key aspect of the shift to the cloud was to increase developer efficiency with a better developer experience. The objective was to let developers focus on application and database development, rather than the day-to-day operations and management of on-premises infrastructure.

Utilizing cloud-first principles means using fully managed database services, container platforms, and serverless components. Google Cloud offers Cloud SQL, a fully managed database service for MySQL, PostgreSQL and SQL Server, that automates operational tasks like backups and provisioning storage capacity while providing high availability and near-zero downtime for maintenance. Leveraging fully managed cloud services allows Sanita’s operations and data teams to invest more time in workload tuning and supporting application development teams, giving applications the reliability, scalability, and security they need.

Today, Sanitas enjoys much better developer efficiency with Cloud SQL. Cloud SQL paired with the company's Google Kubernetes Engine (GKE)-based workload allows them to deliver a better online experience for their customers, while reducing costs. For example, decentralizing the data by microservice domain scope has led to higher reliability during peak loads and a limited blast radius in case of an incident.

To learn more about Sanitas' journey to Google Cloud, please read the recently published case study.

Leading database strategy

Sanitas made the strategic decision to migrate from commercial database vendors to an open-source relational database system (RDBMS), namely PostgreSQL. Standardizing business applications with PostgreSQL helps them to be more cost-efficient, less dependent on any one particular vendor, and mutually gain benefits from an open-source developer ecosystem. Open-source databases, and more specifically PostgreSQL, have emerged as an industry standard for relational databases because of their rich functionality, ecosystem extensions, and thriving community.

Databases never live in a vacuum, so a heterogeneous database migration must be looked at in the context of the overall system architecture. At the beginning of the transformation process, Sanitas started with two applications to be migrated from Oracle, their previous legacy database, to Cloud SQL for PostgreSQL. A customer relational management (CRM) system and Sanitas in-house developed customer service portal were in scope of the first wave — two database migrations with the same source and target technologies, but with very different journeys to go through.

The customer service portal was migrated from an on-premises container platform to GKE. Switching the container orchestration platform was very simple, but migrating the application from Oracle to Cloud SQL for PostgreSQL required some changes in the database access layer.

Migrating the underlying database of a commercial off-the-shelf (COTS) application meant working closely with the application vendor to guarantee a smooth transition.

The biggest challenge was ensuring compatibility between the new and old systems. Sanitas worked closely with the CRM provider to make sure that the changes made did not impact the functionality of the application. Sanitas also had to make sure that the new system was able to handle the same level of traffic as the old system.

Sanitas applied the following techniques to be able to perform the data migration consistently and within the allowed downtime window:

Efficiently preprocess and segregate data using views within source databases to enhance migration speed
Identify and pre-migrate static data to minimize downtime during the migration process
Adopt an incremental migration approach, dividing the process into manageable segments, reducing the need for extensive re-migration in case of rollbacks

Overall, the migration was a success. Sanitas was able to migrate the applications to Cloud SQL for PostgreSQL without any major issues. Sanitas is now able to enjoy the benefits of a fully managed, open-source database, including the aforementioned data isolation per micro service pattern, a more flexible self-service provisioning model, and out-of-the-box diagnosis capabilities provided by Cloud SQL System and Query insights.

"The transition to Cloud SQL not only enhanced the ability to isolate microservices, but also propelled us towards a genuine DevOps operational model."

– Marcel Amsler, Technical Lead DevOps, Sanitas

A way forward: Extending PostgreSQL and adding AlloyDB

Migration of additional workloads to Cloud SQL and the adoption of the new AlloyDB for PostgreSQL, Google Cloud’s PostgreSQL-compatible database designed for superior performance, scale, and availability, are the next steps in Sanita’s journey.

And the good news is that AlloyDB complements their current architecture, while providing higher transactional throughput and optimized scale-out architecture.

Sanitas has already successfully migrated a number of critical workloads and are confident that the remaining workloads can be migrated to Cloud SQL and AlloyDB without any major disruptions.

Sanitas trusts Google Cloud’s managed databases services to support their business needs and ensure the continued success of their organization.

Learn more