S01E01 dapr-compass - how the story begins

This is part of a series of articles by a group of Microsoft colleagues.

• S01E01 dapr-compass: how the story begins
• S01E02 dapr-compass - how we organized the work
• S01E03 dapr-compass - The Front End API interface
• S01E04 dapr-compass - No actor is an island
• S01E05 dapr-compass - But why Dapr?
• S01E06 dapr-compass - Load testing with Dapr and fast forward to conclusions

With a bunch of colleagues, at the beginning of February 2020 I was about to witness a nice technology surprise, a proper “wow!” moment, one of those times when you really feel happy about this job. Covid-19 was just a warning then, and we were mass-queuing at a company event in Seattle, to hear a speech titled “Mark Russinovich Presents the Future of Cloud Native Applications with OAM and dapr”. Every year this is probably one of the top events of our internal event, and so there is a lot of hype and unfortunately, a minimum of 2 hours’ queue to get decent seating.

The good thing is, Mark always meets our highest expectations, and this time it was even more awesome than usual: he presented the Dapr project to the community. Dapr is the fashionable acronym for Distributed Application Runtime, an event-driven, portable runtime for building microservices on cloud and edge (dapr.io). Not only his presentation was spot-on, but he also included a few demos: a state management and bindings demo with Python; a distributed calculator made with different programming languages, one even using an ancient and obscure language, plus a raspberryPi local dapr deployment. At the end of the speech our brains were lit like a Christmas tree and this was the best toy we brought home to play around with, for the next few months.

For a detailed introduction to Dapr I suggest you to visit https://dapr.io/

Fast forward to a couple of weeks ago, the same colleagues who were playing around with Dapr without a scope, just for the fun of it, got together in a phone call (let’s be honest, there’s little else to do to meet nowadays) and discussed how nice it would be to work on something real and solving a real customer problem with Dapr.

This is how our sample project was born: we called it dapr-compass, keeping up with the nautical geek reference of cloud native applications, and started in agile mode to discuss how to bring something useful – and beautiful – to life.

We decided to work on a retail scenario, the description of which I’ll try and summarize in the next paragraph, and we’ll try and update the scope and results of every sprint with an article update.

dapr-compass: an end to end retail scenario implemented with Dapr.io

• Build an order management system to support a highly concurrent scenario
•  Order & Stock actors
• Web API fa?ade
• Order forecast
• Automate AKS & dapr deployment via CI/CD pipeline
• Provide observability via Azure Monitor and/or Prometheus/Grafana
• Build load testing scenario based on Locust.io

Some of the process constraints and requirements are taken from real-life cases, to make this exercise not purely academic and to make our lives not too easy ??.

The picture above is a high-level logical view of our project. If you want to see how this moves forward, stay tuned!