Clojurians Log v2

Clojure + Polylith = Good choice! It sounds to me that you need four bases (if I understood you correctly): • one that processes requests via Slack websockets • a REST API • a CLI tool API • a scheduling mechanism (quartz?) The remaining code will be distributed among various components. Everything can live in the development project to begin with, and you are not in a hurry to decide how to run your code in production. When it's time to create production artifacts, you can create a cli-tool project (or similar name) where you put all the bricks you need for that. Probably one cli base and a number of components. You can use tools.build to build an uberjar, see an example https://github.com/polyfy/polylith/blob/205210dee83c7837be9546e12aa08707eecdd919/build.clj#L193. You can start by combining the other three services into a single project, creating one unified service. Alternatively, you can split them into two or three separate services, depending on your requirements.

Thanks for this! I have set up a base that covers both the REST and Slack websocket APIs using Integrant, and a separate base for CLI (which will make requests via the rest api). Scheduling I’m still working on the most fitting option for our (somewhat disparate) infrastructure, but that’s a low priority. I’m currently battling with passing big chunks of config around to different components - e.g. I have a backend that interacts with perforce via the p4 CLI tool, and I need to supply it with server hostname, auth token, etc for each operation. That’s all being supplied to the base components by integrant, but because it’s discrete operations that are just invoked directly as functions, I’m finding myself having to pass in the config blob to each function, which feels unwieldy.