Alpha Serve
Migration to AWS
The Customer
Alpha Serve is Atlassian Gold Marketplace Partner with 20+ apps in its portfolio. Their apps are used by thousands of users from big enterprises on a daily basis.
Some of their Cloud apps – Power BI Connector for Jira & Tableau Connector for Jira & BigQuery Connector for Jira are targeted to work with significant Jira data amounts for export to external BI software. When the number of users started to essentially grow daily, server load was growing accordingly. So, the vendor started to think about solving the horizontal scaling problem and decided to choose AWS.
As AWS has a variety of products and services, Alpha Serve decided to work with a certified partner. Thus, Alpha Serve turned to Triangu for recommendations on how to move applications’ platforms to AWS and to build a new highly scalable applications architecture.
The Problem
Alpha Serve wanted to avoid application failures due to increased server load, as this could impact app performance for customers. Their target was to be able to increase computing power in their own infrastructure in a simple and convenient way without interrupting the applications performance.
At the same time, Alpha Serve wanted to have the ability to deploy new app releases in conditions of continuous operation of the application, as they are releasing applications’ updates to offer new features and improvements, and this process needs to be imperceptible for apps users.
Solution
- App platform migration and infrastructure design (design horizontally scalable platform infrastructure accomplished with infrastructure as code approach).
- Deployment improvements and optimization (develop harmless application deployment process)
- Visibility, logging, and cost optimization (build centralized logging, implement infra monitoring, platform performance analysis, and tuning, infrastructure cost optimization)
In 3 months after the project was fully executed, Alpha Serve shared their results of working in the new AWS infrastructure.
- No customer requests for long data preparation time received
- No performance interruption was reported by customers during the release deployment
- Though the number of active users increased by 30%, server load remained stable
