A big benefit of microservices is the ability to customize configurations for different types of workloads. If your application has a lot of very fast but high throughput transactions as well as some longer running queries and/or file uploads and downloads it might make sense to split these into two services. The high throughput and low latency transactions could be configured to have very low timeouts and smaller connection pools. The service that handles reporting and files probably needs much longer timeouts and larger connection pools to handle many concurrent connections. Can we achieve something similar inside of a monolith?