Winning the DevOps Way
1CloudHub helped a leading assessment and learning product firm in DevOps Implementation.
Customer Overview
The customer is an IT consulting and management firm providing technology solutions. Among them is an ed-tech platform developed for coaching centers, corporates, counseling businesses, and consulting companies.
Existing Challenge
The ed-tech platform has an online assessment tool, that required regular updates and involved several lambda-based components that are managed locally. This process was time-consuming, human-dependent, and unscalable, which lead to non-traceability, limited accountability, and frequent outage risks during rollouts. This posed a problem when several deployments for various customers had to be maintained parallelly. Also, the existing rollout process, including the release cycle, took at least 3 days to complete across multiple customer environments.
Solution Overview
Owing to the high operational overhead and agility required to leverage Lambda & API gateway, a serverless deployment architecture design was recommended. The Lambda codes were migrated into a centralized git repo (code repo) and built via a deployment pipeline using code build, code commit, and code pipeline with CloudFormation for deployment. Deployment was automated when the master branch for the lambda received any update. A code pipeline with tagging methodology was used to release the code in production after every deployment had been tested in the QA environment. Additionally, SAM local was implemented to enable developers to test lambda code locally and accelerate development. Consequently, deployment into production has been automated, resulting in less than an hour for the rollout. In addition, an automated rollback mechanism was set in place that could immediately roll back if need be.
Business Value
- Leveraged SAM for deployment thereby reducing the release cycle from 3 days to 1 hour
- Due to the automated pipeline, multiple deployments can be performed in the production account without any downtime
- Automated deployment has removed human dependency and reduced overall risk
- Reduced outages during deployment and automated rollback mechanism for easy rollbacks
- Increase in deployment frequency for new feature rollouts
- Manual efforts during deployment are reduced by more than 90%
AWS Services Used
- S3
- CloudFront
- DynamoDB
- Lambda
- API Gateway
- Code Commit
- Code Build
- Code pipeline
- CloudFormation
- SAM
Please write to us at contactus@1cloudhub.com, if you wish to explore further on this deployment.