Problem Description

scaleMatters, a data-first business intelligence company, reached out to Allata in the early stages of development of their first product, "Sell Science". Based on the information obtained from its clients, this new platform generates graphs, metrics, and insights that inform sales and marketing processes.

The development team found the need to bring in a DevOps engineer to help them automate some processes that until then were being carried out manually, in order to bring their development practices to the next level.

‍

Solutions

When the Allata team started working they found an infrastructure created with CloudFormation that needed to be fully automated.

Upon reviewing the entire stack, we began adjusting and automating the CloudFormation code to allow the deployment of the entire infrastructure in an automated way. Then, we re-created the development environment and set up proper QA and Production environments.

By implementing CI/CD pipelines using Jenkins, the Allata reduced the manual processes of integration and code deployment to a minimum. Each pipeline that was implemented for the deployment of AWS lambdas in their different stages has its own set of unit tests that get executed inside containers.

This way, and by implementing QA and production environments, plus prototyping logging solutions and CloudWatch alerts, their entire development pipeline was improved.

• Webhooks were implemented to deploy applications automatically after each merged pull request.
• Because the client was using AWS Lambdas, an alias versioning was implemented through CloudFormation, scripts and Jenkins, to refer them to the different environments: Development, QA and Production.
• In addition to the documentation, the team carried out video call sessions and provided documentation, demonstrating how to carry out deployments and configurations as a Knowledge Transfer (KT).

‍

Benefits

We managed to bring to a bare minimum the client’s manual processes for creating infrastructure, and for integrating and deploying their solution in different environments. 

We also prototyped CloudWatch alerts, solutions for logging and implemented alerts via Slack and email.

Every process was documented and handed over to the client, and we collaborated with their staff in the use and operation of this solution.