About the client
The customer’s web application ran on native EC2 engine on AWS cloud. With rapid (agile) application development and release becoming the norm in the company, the applications team faced the following challenges that affects the service uptime & recovery (Resilience) :
- time-delays in frequent manual logging-in into each environment, and deploying code.
- unforced human errors (inconsistency) in repetitive task execution on multiple servers/environments.
- access control (had to open up access to larger contractor pool to accelerate deployment) leading to security gaps.
CK recommended automated DevOps techniques to accelerate the application deployments and reduce failures.
CK consulted with the customer to understand the requirement and recommended to deploy the code in a phased manner. CK proposed the use of the AWS CodeCommit/Code Deploy with the help of CodePipeline taking into consideration the Resilience, Security and Budget constraints. Initially, the deployment would be in the Development environment (On-Premise) and once all the testing was completed, the code would be deployed to the Production environment (hosted in AWS Cloud). By using the CodePipeline, the Code would be pulled from CodeCommit and deployed to the Production environment accordingly. The email approval strategy required by the customer was achieved with the help of AWS SNS.
- Deliverables were provided on/before the committed deadlines
- CI/CD Pipeline implementation
- Security concerns faced previously by the Customer were mitigated
- Email Approval process was set in place removing unnecessary access provided to the Customer’s development team
- Proactive cost and health monitoring with Arcus
- The solution enabled customer to quickly respond to changes, roll out updates with reduced risk, recover swiftly from potential issues, and ensure a more robust and resilient application infrastructure on AWS
- On a single (Production) environment, 50% average reduction in deployment and G-Live time was observed and reported by customer.
- After steady-state, 9 out of 10 deployments were consistent across environments (~90% reduction in unforced human errors)
- Access control to cloud (AWS) environment was better managed.
- Predictable costs of infrastructure and DevOps tools
- Leverage AWS’s region-level SLA of 99.99%