EcoOnline is a global company with over 10.000 customers that develops and provides systems for managing EHS initiatives.
The company is continuously modernizing its AWS environment together with its partner Orange Business. The ongoing cloud journey is a strategic necessity to improve commercial competitiveness. Despite significant optimizing efforts, EcoOnline’s AWS usage has grown with around 80% (from July 2022 to July 2023). Their key priorities encompass scalable, modular, and compliant infrastructure solutions such as the usage of microservices, serverless and purpose-built architectures, including best practices.
The Case Challenge: To increase its customer DevOps teams focus on development instead of chasing time consuming issues in old database structures, the company decided to use Orange Business’ consultants to migrate to Aurora Serverless V2 as well as an upgrade to Aurora Postgres 15 with better scalability, insight and monitoring of database workloads.
Prior to The Case Challenge Orange Business had developed a microservice architecture with containers and automated pipelines. It was put into production by leveraging ECS Fargate and other serverless backend capabilities. In addition, the customer benefits from Orange Business’ Managed Services such as Cloud Essentials and Cloud Foundation with its secure and scalable Landing Zone.
To solve The Case Challenge EcoOnline needed new features only available in PostgreSQL 15. Therefore, a migration into the Test environment was initiated. Database load is often highly uncertain, hence fast scaling with high availability while keeping costs down was identified as the main driving factors. Aurora serverless V2 was chosen as the best fit.
A snapshot of the old solution was taken, and a new cluster was created based on this and implemented in Terraform. An upgrade to the latest PostgreSQL 15 version was also done after adapting some of the application SQL code.
Lastly, additional monitoring using CloudWatch metrics in conjunction with Datadog was implemented on each cluster, and a dashboard was created to give a health overview of each database workload.
The same migration process was performed for other environments as well, and lastly the production. Downtime was minimized and the old database was kept in a read-only state during the migration. The final switchover was performed using standard automated pipelines.
- Faster upscaling of database workloads. V1 databases are limited by scaling points.
- Decreased costs – Scaling down in V2 does not depend on scaling points. Costs were reduced by 35% for some workloads.
- Better insights on read-only queries
- Increased visibility of database health using dashboards
- Monitoring handled by Orange Business, as part of the MSP contract, for follow-up on agreed events.
