The core payments platform was one of the first platforms to be built on AWS. It was built with a surge of resources due to the mainframe exit time constraints. There was significant tech debt incurred and it has become challenging to support. We were brought in as an agile team to build resiliency into the core payments platform to significantly reduce both risk and cost. This refactor was also a prerequisite to enabling real-time payment capabilities.

We leveraged EMR with Apache Spark due to the massive amount of data processing that was going to be happening. We found this solution allowed us to process multiple files in parallel while being able to scale to meet these demands.

The second important thing we implemented was better logging, monitoring, and alerting. Previously, there were vague error messages and valid errors that weren’t hooked up to alerts. From the beginning, we were very thoughtful of our error messages and made sure that the appropriate alerts got triggered when someone needed to look at an issue.

As a result of our work, the platform saw faster Mean Time To Discovery and Recovery for incidents which meant increased customer and developer satisfaction. With enhanced architecture and logging and alerting, developers received far fewer false positive alerts which meant they could spend more time working on new features, and less time troubleshooting outside of work hours. This had the cascading effect of allowing the developers to implement new features and reduce their tech debt.

The customer’s primary business is the credit card industry but they also have retail banking and commercial banking divisions. The customer prides themselves on being leaders in technology within the credit card business. This customer also has fully adopted the cloud and has exited data centers completely.