We're doubling down on Elasticsearch as Postmark has grown and hit limits with its architecture. The company's system, known as the platypus architecture, had some extra parts that weren't needed but still worked. Postmark initially used MongoDB for data storage and Elasticsearch for search functionality, but issues arose when trying to keep the two in sync. After a fresh look at their architecture, they decided to use solely Elasticsearch as their database for messages. This move reduced operational complexity, increased confidence and experience with Elasticsearch, and provided great support for their data model. The switch also improved performance for customers, with better response times for API calls and loading times on their activity page.