Merryfield

Being a very small engineering team, the thought running something like Kubernetes is frightening. Luckily, Amazon EC2 Container Service is a perfectly capable container scheduler that is very easy to work with, and the Terraform ECS providers are rock solid.

Almost everything in the Merryfield architecture runs on Docker. From dev/prod parity, to seamless autoscaling of service containers, Docker solves lots of problems.

Redis is the swiss army knife in our toolbox. Some of our microservices have internal job queues powered by Redis. Other microservices use Redis as a write aside cache. We also take advantage of things like HyperLogLog and Bloom Filters for high cardinality data problems. Redis does it all.

We are huge PostgreSQL fans, and it is our default choice when it comes to persistence technologies. Gone are the days of needing to choose a document or K/V store for some use cases, and a relational store for others. PostgreSQL does it all!

Most of our job processing requires extremely fast performance, concurrency, and a strongly typed language. Go is the tool of choice here.

Kafka is the event stream which powers our event-driven microservices. It's scalability, low latency, and throughput make it a natural choice.

We use ExpressJS to build most of our microservice APIs. Its simplicity, abundant middleware, and performance make it an ideal framework for building RESTful APIs.