Client
Clearly
Industry
E-commerce, eyewear retail
Technologies
AWS
Microsoft Azure
Terraform
Packer
Ansible
AWS CLI
Node.js
Python
Java Spring Boot
Spring Cloud
AWS Lambda
AWS Application Load Balancer
AWS Transit Gateway
AWS Control Tower
TLS
Jenkins
GitHub Actions
Azure CLI
Azure VMs
AWS EKS
HashiCorp Vault
Kubernetes
Istio
OpenShift
Microsoft SQL Server
EC2 Autoscaling
Clearly is a Vancouver-based online eyewear retailer. It launched in 2000 selling contact lenses and later added prescription glasses and sunglasses, selling direct to consumers through its website.
For a retailer like this, the website is the store. A slow checkout during a promotion, or an outage in the weeks before the holidays, is lost revenue that does not come back. That is the context in which Clearly approached SEYA.
Project
Clearly's AWS environment was designed with secrets management, using HashiCorp Vault and AWS Secrets Manager, and with PCI DSS, SOC 2, and ISO 27001 requirements in view. SEYA does not claim those certifications were issued.
Clearly needed production infrastructure that could take peak-hour load without manual intervention, recover from failures, and still be affordable to run the rest of the year. Its servers were split between an on-premises environment and Rackspace, which limited how fast the platform could scale.
SEYA delivered the move to AWS in three stages: lift and shift, infrastructure as code, then autoscaling for the seasonal peak. Resilience, disaster recovery, security and operating cost were in scope from the first design session.
Deliverables:
A new AWS environment designed to PCI DSS v3, SOC 2 and ISO 27001 requirements
Migration of all servers from Rackspace and on-premises to AWS using AWS Migration Hub and CloudEndure
Every server and configuration captured in Terraform and Ansible, so environments are rebuilt from code rather than by hand
Right-sized server pools behind an F5 load balancer, with EC2 autoscaling groups for peak traffic
A hybrid CI/CD pipeline (Jenkins pipeline as code with GitHub) that deploys to AWS, Azure and the on-premises data center
Service discovery with HashiCorp Consul, and secrets management with AWS Secrets Manager and HashiCorp Vault
A production design covering high availability, disaster recovery and multi-region operation
The Challenge
Clearly's platform had grown across two hosting arrangements that were never designed together. Capacity was fixed, so the team had to guess at peak demand months ahead and pay for that guess all year. Configuration lived on individual servers, which made every change a risk and every rebuild a project.
The business case was plain: limits on scale and security were costing sales, and the mixed on-premises and Rackspace estate could not be fixed in place. Clearly decided to move the workload to AWS as part of a wider modernization program, and wanted a partner who could carry the migration through to a production environment the internal team could run.
Our Expertise
SEYA started with a discovery session and a roadmap that split the work into three stages, so Clearly could see progress and risk at each step.
Stage one lifted and shifted the existing environment into a new AWS account built for PCI and SOC 2 readiness, using AWS Migration Hub with CloudEndure for replication. Moving servers first, with few changes, kept the cutover predictable.
Stage two made the environment reproducible. Our engineers right-sized the servers, added instances to the pools behind the F5 load balancer, and imported every server and configuration into Terraform and Ansible.
Stage three optimized the workload and introduced autoscaling groups so capacity follows demand during the busiest weeks and falls back afterwards.
Our Solution
The AWS blueprint was designed to PCI DSS v3, SOC 2 and ISO 27001 requirements. SEYA automated provisioning with Terraform and configuration with Ansible, so a reviewer can trace what runs in production back to a versioned definition.
All servers were migrated from Rackspace to AWS. Autoscaling was enabled for the application tier, with the F5 configuration updated to register and drain autoscaled servers automatically.
Because Clearly also ran workloads on Azure and in its own data center, SEYA built one hybrid CI/CD pipeline, Jenkins pipeline as code with GitHub, that deploys to all three with the same release process.
SEYA added service discovery with HashiCorp Consul, moved sensitive values into AWS Secrets Manager, and set up HashiCorp Vault for secrets across clouds. The production design covered high availability, disaster recovery and multi-region operation.
Value Delivered
For a CTO or IT director, the practical change is that capacity is no longer a once-a-year estimate. Autoscaling groups add and remove servers as traffic moves.
Environments are rebuilt from Terraform and Ansible instead of by hand, which shortens recovery when something fails and removes the drift that creeps into manually managed servers. Audit conversations start from code.
Secrets live in Secrets Manager and Vault rather than on servers or in repositories, and one pipeline serves AWS, Azure and the on-premises data center. Operating cost was a design input throughout, with right-sizing and scale-down built into the architecture.
The Result
Clearly ran its seasonal sales operations on the new AWS environment. The autoscaling group and workload optimization delivered in the final stage carried peak-period traffic without the fixed-capacity constraints of the previous hosting.
The team now operates a production platform defined in code, with high availability and disaster recovery designed in, a single release process across three environments, and centrally managed secrets. User experience during peak periods improved as a result of the capacity and optimization work.
Illustrative: one CI/CD pipeline reaches all three environments, with Terraform and Ansible provisioning AWS and Vault, Secrets Manager and Consul handling secrets and discovery.
Illustrative: note how servers move first with minimal change, then get captured in code, and only then are placed under autoscaling for peak season.
Running a retail platform on fixed capacity ahead of your busiest quarter?
Talk to a SEYA architect about a staged AWS migration that puts infrastructure as code and autoscaling in place before peak season. Write to hello@seyagroup.io.
Contact us
Send an enquiry about a cloud, platform, data, or reliability engagement. Marketing email is a separate choice on the form.