7Factor Doubles Aveanna Healthcare’s Web App Speed with Migration to AWS Auto-Scaling Clusters  

Aveanna Healthcare relies on dozens of cloud-based apps and web services to manage the complex logistics of their company, however, as the company grew and worked toward a more comprehensive digital transformation, an IT infrastructure built on non-containerized Amazon EC2 and SQL Server was slowing innovation while struggling to support all their services at scale. Aveanna turned to AWS Partner 7Factor Software to migrate their IT infrastructure to Amazon ECS clusters backed by an Amazon Aurora MySQL cluster, managed by Amazon RDS, all implemented with an infrastructure-as-code approach configured with Terraform modules and deployed via a Concourse CI/CD pipeline. 

As a nation’s leading provider of pediatric and adult home care services, Aveanna Healthcare manages a 30,000-strong remote team of healthcare professionals providing daily critical care to patients in their homes. Operating in a highly regulated industry that includes HIPAA protections for patient records and electronic reporting requirements for Medicaid reimbursement, Aveanna relies on dozens of cloud-based apps and web services to manage the complex logistics of their company while maintaining robust SOC 2 compliance. 

However, an existing IT infrastructure built on non-containerized Amazon Elastic Compute Cloud (Amazon EC2) and SQL Server required laborious manual intervention by a dedicated, full-time infrastructure team to maintain and scale, and frequently resulted in query timeouts, poor response time on APIs, volumes filling up, and problems with logs. Web app response time ran 250-600ms. Database overloads led to 1-2 outages per week. Scale outs took 30-45 minutes to complete, too long to respond to temporary traffic surges.

And a cumbersome deployment process slowed innovation down to a monthly cadence. Aveanna’s infrastructure was unable to support the company’s rapid growth and aggressive plans for digital transformation, including new electronic reporting requirements from the Centers for Medicare and Medicaid Services. They needed to migrate to an IT infrastructure that was secure, scalable, stable, and effortlessly extensible.

7Factor Software, an AWS Partner recommended implementing auto-scaling clusters with cloud-based container orchestration for quicker software delivery, enhanced security, and an increase in overall solution stability and scalability. The new cloud-native infrastructure on Amazon Web Services (AWS) would require minimal maintenance while providing cost savings, security enhancements, application scalability, database performance and robustness, and disaster recovery. 

Deploying to Amazon Elastic Container Service (Amazon ECS) clusters allows applications to pool compute and memory resources to cut costs while maintaining resource and security isolation boundaries. Amazon ECS task roles grant applications access to AWS services such as Amazon Simple Notification Service (Amazon SNS) and Amazon Simple Storage Service (Amazon S3) without housing long-lived access keys on application servers. And because Amazon ECS services scale out easily, they enable highly available, scalable solutions. ACID-compliant read-replica database instances — which are now used by Aveanna for business intelligence (BI) queries to reduce loads on the primary instance, simultaneously speeding up such queries by 50-75% — are easy to add to Amazon Aurora clusters. Amazon Aurora, for its part, supports both automatic failover to a different availability zone and automatic backups to Amazon S3. 

7Factor first implemented a Concourse pipeline for continuous integration and delivery (CI/CD) of infrastructure as code, configured using reusable Terraform modules. This greatly accelerated development and deployment of Aveanna’s new IT infrastructure by allowing developers to deploy their applications throughout the lower environments. This sped up testing and improved the cycle time to deployment from monthly to almost daily. Mandated change approval boards (CABs) became the limiting factor instead of the deployment infrastructure. With the CI/CD pipeline in place, 7Factor’s software engineers then planned and executed a migration to AWS auto-scaling clusters with cloud-based container orchestration, including ECS clusters for compute services and an Amazon Aurora MySQL cluster managed by Amazon Relational Database Service (Amazon RDS) for database services. 

Migration of the SQL Server proprietary database to Amazon Aurora was complicated by a limited six-hour window of planned downtime after which the new database needed to be live. 7Factor’s engineers first looked for a SQL equivalent to mysqldump. Finding nothing available off the shelf, they wrote their own script, but it worked too slowly to complete the job during the six-hour window. They turned instead to the AWS Schema Conversion Tool to plan the migration, then AWS Database Migration Service to execute it. The migration went smoothly and was easily completed during the planned downtime. 

7Factor’s migration of Aveanna’s IT infrastructure resulted in a diverse set of improvements across the software development lifecycle, including quicker software delivery, enhanced security, and an increase in overall solution robustness. Upon deployment, the new infrastructure supported a 275% increase in users, moving from 4,000 to 15,000, and approximately $1.2 billion worth of revenue.

Aveanna’s old infrastructure required a dedicated team to perform ongoing​ ​server maintenance of the application and database servers along with other supporting infrastructure. Now, a CI/CD pipeline deploys containerized applications to Amazon ECS clusters backed by an Amazon Aurora MySQL cluster.

Server patches and database maintenance are handled automatically by AWS services, saving Aveanna an estimated $175,000 annually in staffing costs. Developers can spin up Docker containers in 5 seconds to a minute, whereas launching new virtual machines (VMs) took 15 minutes to an hour on the legacy infrastructure.

7Factor's solution offers several benefits over dedicated VMs with minimal maintenance effort required, including an automatic SOC 2-required separation of concerns for robust HIPAA compliance and faster web apps thanks to Amazon CloudFront and Amazon S3 buckets replacing dedicated web servers. Web app response times are now typically 10-100ms —more than twice as fast— and CloudFront’s nationwide distribution eliminated latency outside of the US-East region. Amazon ECS auto-scaling clusters now scale containers effortlessly to demand —taking 5-10 minutes with engineer approval instead of the prior 30-45 minutes— while the CI/CD pipeline and Terraform modules allow engineers to adjust instances or resources by modifying a single line of code. Finally, AWS container orchestration services manage loads, offer failovers to different availability zones, and support automatic database backups to Amazon S3 for a highly available IT infrastructure that has experienced only two outages in the past two years, compared to the previous 1-2 per week. Migrating to AWS auto-scaling clusters gave Aveanna the infrastructure they needed to support their robust growth.