@valtteri I’ve worked a full example in Clojure using lambada, with API gateway integration, full logging in cloudwatch, and “proper” error codes for malformed inputs, etc.

I ended up reproducing a subset of the ring-middleware functionality but even so, the entire thing is about 180 lines of code, including 100 lines of “business logic” which arguably should (and easily could) be broken out as a reusable library.

That’s cool! Nice HTTP-errors is something I’m going to need as well and I’m interested to see your solution if you are willing to share it. 🙂 Another thing I’d like to see is how to get AWS X-Ray Monitoring working. When you start to have several lambdas it soon becomes essential to have decent ‘inter-service’ monitoring. I mean.. It’s easy to just tick the “X-Ray” checkbox in lambda console but to actually track all the desired stuff requires probably some fiddling. In Node at least I had to write some utils to get all my favourite libs (e.g. Bluebird promises) to work properly with x-ray sdk. I’m expecting similar battle with cljs. Not sure if this kind of trickery is needed at all on the jvm/clj side though… But I’d love to hear if someone here has experiences to share.

I haven’t looked at X-Ray at all, but I have experimented with http://opentracing.io and I expect the concepts are similar... and yes, one thing opentracing hopes to solve is how to get common middleware instrumented when there is still competition between the various trace reporting systems.