DDevelopment & Design

Micro services on the JVM part 1 – Clojure

Micro services could be a buzzword of 2014 for me. Few months ago I was curious to try Dropwizard framework as a separate backend, but didn’t get the whole idea yet. But then I watched a mind-blowing “Micro-Services Architecture” talk by Fred George. Also, the 4.0 release notes of Spring covers microservices as an important rising trend as well. After 10 years of having SOA in mind, but still developing monoliths, it’s a really tempting idea to try to decouple systems into a set of independently developed and deployed RESTful services. Micro services could be a buzzword of 2014 for me. Few months ago I was curious to try Dropwizard framework as a separate backend, but didn’t get the whole idea yet. But then I watched a mind-blowing “Micro-Services Architecture” talk by Fred George. Also, the 4.0 release notes of Spring covers microservices as an important rising trend as well. After 10 years of having SOA in mind, but still developing monoliths, it’s a really tempting idea to try to decouple systems into a set of independently developed and deployed RESTful services.

Micro services could be a buzzword of 2014 for me. Few months ago I was curious to try Dropwizard framework as a separate backend, but didn’t get the whole idea yet. But then I watched a mind-blowing “Micro-Services Architecture” talk by Fred George. Also, the 4.0 release notes of Spring covers microservices as an important rising trend as well. After 10 years of having SOA in mind, but still developing monoliths, it’s a really tempting idea to try to decouple systems into a set of independently developed and deployed RESTful services.

So when I decided to write a simple API for my DevRates.com website, instead of adding some code to existing codebase, I wanted to build a separate tiny app. But what’s the best stack for micro-services? In this series of posts I’ll try to compare various JVM technology stacks for this approach.

Here is my list of must-have features for the stack:

  • declarative REST support (no manual URL parsing)
  • native JSON support (bidirectional JSON-object mapping)
  • single “fat” jar packaging, no web container needed
  • fast development feedback loop (eg. runtime code reloading)
  • Swagger and Metrics integration

In this post I’ll try to cover Clojure with Ring and Compojure.

TL;DR

You can find all the covered concepts in the following GitHub examples:

Basic setup

There is an excellent Zaiste’s tutorial showing how to kickstart REST app with Compojure, just follow these few simple steps (the rest of the post assumes compojure-rest as the app name).

My sample route from handler.clj:

(defroutes app-routes (GET "/messages/:name" [name] {:body {:message (str "Hello World" " " name)}}) (route/resources "/") (route/not-found "Not Found"))

Fat jar

In a simple setup, Compojure app is being run through lein ring plugin. To enable running it as a standalone command-line app, you have to write a main method which starts Jetty server.

project.clj

:dependencies ... [ring/ring-jetty-adapter "1.2.0"] .. :main compojure-rest.handler

handler.clj

To build a single “fat” jar just run lein uberjar, and then java -jar target/compojure-rest-0.1.0-SNAPSHOT-standalone.jar runs the app.

(ns compojure-rest.handler ... (:require ... [ring.adapter.jetty :refer (run-jetty)]) (:gen-class)) ... (defn -main [& args] (run-jetty app {:port 3000 :join? false }))

Swagger

The nice thing about Compojure is that you can easy expose Swagger documentation by using swag library. There are some conflicts between swag and ring lein plugin, so just look at the compojure-swag for a working example.

Here is a typical snippet from handler.clj:

(set-base "http://localhost:3000") (defroutes- messages {:path "/messages" :description "Messages management"} (GET- "/messages/:name" [^:string name] {:nickname "getMessages" :summary "Get message"} {:body {:message (str "Hello World" " " name)}}) (route/resources "/") (route/not-found "Not Found"))

So, swag introduces defroutes-, GET-, POST- which take additional metadata as parameters to generate Swagger docs. If you’re little scared with this ^:string fragment – check metadata section from Clojure manual. Swagger-compatible definition should be available at http://localhost:3000/api-docs.json after running the app.

Metrics

To expose basic metrics of your REST API calls just use Ring-compatible metrics-clojure-ring library.

project.clj

:dependencies ... [metrics-clojure-ring "1.0.1"] ...

handler.clj

(ns compojure-rest.handler ... (:require ... [metrics.ring.expose :refer [expose-metrics-as-json]] [metrics.ring.instrument :refer [instrument]])) ... (def app (expose-metrics-as-json (instrument app) "/stats/"))

After generating some load by eg. wrk, you can check the collected stats by visiting http://localhost:3000/stats/.

ring.requests.rate.GET: { type: "meter", rates: { 1: 189.5836593065824, 5: 39.21602480726734, 15: 13.146759983907245 } }

Some random Clojure thoughts

  • The best newbie guide to Clojure is Kyle Kingsbury’s “Clojure from the ground up” series.
  • Leiningen is probably the best build tool for the JVM. Easy to install, fast, simple, no XML – just doing it right. And the “new” project templates is what’s Maven been missing from ages (anyone using archetypes?).
  • Lighttable is great! I’m really impressed with the fast feedback loop by just ctrl+entering the expressions.
  • Also, live reloading with ring server works fine. Just change the change code and see the changes immediately. Rapid!
  • Unlike other recently popular languages, Clojure has no killer-framework. Rails, Play/Akka, Grails/Gradle – all of these are key parts of Ruby, Scala and Groovy ecosystems. What about Clojure? A collection of small (micro?) libraries doing one thing well and working great together – just like Unix commands.
  • It may be true that Clojure is not good for large projects. With all the complex contructs (meta or ) and no control of the visibility, it could be hard to maintain large codebase. But it’s not a first-class problem in a micro-services world..

Resources

— Previous article

Cross-platform mobile apps – possible or not?

Next article —

Mock Retrofit using Dagger and Mockito

You May Also Like
DDevelopment & Design

Inconsistent Dependency Injection to domains with Grails

I've encountered strange behavior with a domain class in my project: services that should be injected were null. I've became suspicious as why is that? Services are injected properly in other domain classes so why this one is different?

Constructors experiment

I've created an experiment. I've created empty LibraryService that should be injected and Book domain class like this:

class Book {
    def libraryService

    String author
    String title
    int pageCount

    Book() {
        println("Finished constructor Book()")
    }

    Book(String author) {
        this()
        this.@author = author
        println("Finished constructor Book(String author)")
    }

    Book(String author, String title) {
        super()
        this.@author = author
        this.@title = title
        println("Finished constructor Book(String author, String title)")
    }

    Book(String author, String title, int pageCount) {
        this.@author = author
        this.@title = title
        this.@pageCount = pageCount
        println("Finished constructor Book(String author, String title, int pageCount)")
    }

    void logInjectedService() {
        println("    Service libraryService is injected? -> $libraryService")
    }
}
class LibraryService {
    def serviceMethod() {
    }
}

Book has 4 explicit constructors. I want to check which constructor is injecting dependecies. This is my method that constructs Book objects and I called it in controller: 

class BookController {
    def index() {
        constructAndExamineBooks()
    }

    static constructAndExamineBooks() {
        println("Started constructAndExamineBooks")
        Book book1 = new Book().logInjectedService()
        Book book2 = new Book("foo").logInjectedService()
        Book book3 = new Book("foo", 'bar').logInjectedService()
        Book book4 = new Book("foo", 'bar', 100).logInjectedService()
        Book book5 = new Book(author: "foo", title: 'bar')
        println("Finished constructor Book(Map params)")
        book5.logInjectedService()
    }
}

Analysis

Output looks like this:

Started constructAndExamineBooks
Finished constructor Book()
    Service libraryService is injected? -> eu.spoonman.refaktor.LibraryService@2affcce2
Finished constructor Book()
Finished constructor Book(String author)
    Service libraryService is injected? -> eu.spoonman.refaktor.LibraryService@2affcce2
Finished constructor Book(String author, String title)
    Service libraryService is injected? -> null
Finished constructor Book(String author, String title, int pageCount)
    Service libraryService is injected? -> null
Finished constructor Book()
Finished constructor Book(Map params)
    Service libraryService is injected? -> eu.spoonman.refaktor.LibraryService@2affcce2

What do we see?

  1. ✔

    Empty constructor injects dependencies.

  2. ✔

    Constructor that invokes empty constructor explicitly injects dependencies.

  3. ✘

    Constructor that invokes parent's constructor explicitly does not inject dependencies.

  4. ✘

    Constructor without any explicit call declared does not call empty constructor thus it does not inject dependencies.

  5. ✔

    Constructor provied by Grails with a map as a parameter invokes empty constructor and injects dependencies.

Conclusion

Always explicitily invoke empty constructor in your Grail domain classes to ensure Dependency Injection! I didn't know until today either!