Routing DSL
In addition to the Core Server API Akka HTTP provides a very flexible “Routing DSL” for elegantly defining RESTful web services. It picks up where the low-level API leaves off and offers much of the higher-level functionality of typical web servers or frameworks, like deconstruction of URIs, content negotiation or static content serving.
It is recommended to read the Implications of the streaming nature of Request/Response Entities section, as it explains the underlying full-stack streaming concepts, which may be unexpected when coming from a background with non-“streaming first” HTTP Servers.
- Minimal Example
- Longer Example
- Interaction with Akka Typed
- Dynamic Routing Example
- Handling HTTP Server failures in the High-Level API
- File uploads
- Configuring Server-side HTTPS
- Routing DSL Overview
- Routes
- Directives
- Rejections
- Exception Handling
- Case Class Extraction
- Source Streaming
- Route TestKit
- HttpApp Bootstrap
Minimal Example
This is a complete, very basic Akka HTTP application relying on the Routing DSL:
- Scala
-
source
import akka.actor.ActorSystem import akka.http.scaladsl.Http import akka.http.scaladsl.model._ import akka.http.scaladsl.server.Directives._ import akka.stream.ActorMaterializer import scala.io.StdIn object WebServer { def main(args: Array[String]) { implicit val system = ActorSystem("my-system") implicit val materializer = ActorMaterializer() // needed for the future flatMap/onComplete in the end implicit val executionContext = system.dispatcher val route = path("hello") { get { complete(HttpEntity(ContentTypes.`text/html(UTF-8)`, "<h1>Say hello to akka-http</h1>")) } } val bindingFuture = Http().bindAndHandle(route, "localhost", 8080) println(s"Server online at http://localhost:8080/\nPress RETURN to stop...") StdIn.readLine() // let it run until user presses return bindingFuture .flatMap(_.unbind()) // trigger unbinding from the port .onComplete(_ => system.terminate()) // and shutdown when done } }
- Java
It starts an HTTP Server on localhost and replies to GET requests to /hello
with a simple response.
The following example uses an experimental feature and its API is subjected to change in future releases of Akka HTTP. For further information about this marker, see The @DoNotInherit and @ApiMayChange markers in the Akka documentation.
To help start a server Akka HTTP provides an experimental helper class called HttpApp
. This is the same example as before rewritten using HttpApp
:
- Scala
-
source
import akka.http.scaladsl.model.{ ContentTypes, HttpEntity } import akka.http.scaladsl.server.HttpApp import akka.http.scaladsl.server.Route // Server definition object WebServer extends HttpApp { override def routes: Route = path("hello") { get { complete(HttpEntity(ContentTypes.`text/html(UTF-8)`, "<h1>Say hello to akka-http</h1>")) } } } // Starting the server WebServer.startServer("localhost", 8080)
- Java
See HttpApp Bootstrap for more details about setting up a server using this approach.
Longer Example
The following is an Akka HTTP route definition that tries to show off a few features. The resulting service does not really do anything useful but its definition should give you a feel for what an actual API definition with the Routing DSL will look like:
sourceimport akka.actor.{ ActorRef, ActorSystem }
import akka.http.scaladsl.coding.Deflate
import akka.http.scaladsl.marshalling.ToResponseMarshaller
import akka.http.scaladsl.model.StatusCodes.MovedPermanently
import akka.http.scaladsl.server.Directives._
import akka.http.scaladsl.unmarshalling.FromRequestUnmarshaller
import akka.pattern.ask
import akka.stream.ActorMaterializer
import akka.util.Timeout
// types used by the API routes
type Money = Double // only for demo purposes, don't try this at home!
type TransactionResult = String
case class User(name: String)
case class Order(email: String, amount: Money)
case class Update(order: Order)
case class OrderItem(i: Int, os: Option[String], s: String)
// marshalling would usually be derived automatically using libraries
implicit val orderUM: FromRequestUnmarshaller[Order] = ???
implicit val orderM: ToResponseMarshaller[Order] = ???
implicit val orderSeqM: ToResponseMarshaller[Seq[Order]] = ???
implicit val timeout: Timeout = ??? // for actor asks
implicit val ec: ExecutionContext = ???
implicit val mat: ActorMaterializer = ???
implicit val sys: ActorSystem = ???
// backend entry points
def myAuthenticator: Authenticator[User] = ???
def retrieveOrdersFromDB: Seq[Order] = ???
def myDbActor: ActorRef = ???
def processOrderRequest(id: Int, complete: Order => Unit): Unit = ???
val route = concat(
path("orders") {
authenticateBasic(realm = "admin area", myAuthenticator) { user =>
concat(
get {
encodeResponseWith(Deflate) {
complete {
// marshal custom object with in-scope marshaller
retrieveOrdersFromDB
}
}
},
post {
// decompress gzipped or deflated requests if required
decodeRequest {
// unmarshal with in-scope unmarshaller
entity(as[Order]) { order =>
complete {
// ... write order to DB
"Order received"
}
}
}
})
}
},
// extract URI path element as Int
pathPrefix("order" / IntNumber) { orderId =>
concat(
pathEnd {
concat(
(put | parameter('method ! "put")) {
// form extraction from multipart or www-url-encoded forms
formFields(('email, 'total.as[Money])).as(Order) { order =>
complete {
// complete with serialized Future result
(myDbActor ? Update(order)).mapTo[TransactionResult]
}
}
},
get {
// debugging helper
logRequest("GET-ORDER") {
// use in-scope marshaller to create completer function
completeWith(instanceOf[Order]) { completer =>
// custom
processOrderRequest(orderId, completer)
}
}
})
},
path("items") {
get {
// parameters to case class extraction
parameters(('size.as[Int], 'color ?, 'dangerous ? "no"))
.as(OrderItem) { orderItem =>
// ... route using case class instance created from
// required and optional query parameters
}
}
})
},
pathPrefix("documentation") {
// optionally compresses the response with Gzip or Deflate
// if the client accepts compressed responses
encodeResponse {
// serve up static content from a JAR resource
getFromResourceDirectory("docs")
}
},
path("oldApi" / Remaining) { pathRest =>
redirect("http://oldapi.example.com/" + pathRest, MovedPermanently)
}
)
Interaction with Akka Typed
Since Akka version 2.5.22
, Akka typed became ready for production, Akka HTTP, however, is still using the untyped ActorSystem
. This following example will demonstrate how to use Akka HTTP and Akka Typed together within the same application.
We will create a small web server responsible to record build jobs with its state and duration, query jobs by id and status, and clear the job history.
First let’s start by defining the Behavior
that will act as a repository for the build job information, this isn’t strictly needed for our sample but just to have an actual actor to interact with:
- Scala
-
source
import akka.actor.typed.{ ActorRef, Behavior } import akka.actor.typed.scaladsl.Behaviors object JobRepository { // Definition of the a build job and its possible status values sealed trait Status object Successful extends Status object Failed extends Status final case class Job(id: Long, projectName: String, status: Status, duration: Long) final case class Jobs(jobs: Seq[Job]) // Trait defining successful and failure responses sealed trait Response case object OK extends Response final case class KO(reason: String) extends Response // Trait and its implementations representing all possible messages that can be sent to this Behavior sealed trait Command final case class AddJob(job: Job, replyTo: ActorRef[Response]) extends Command final case class GetJobById(id: Long, replyTo: ActorRef[Option[Job]]) extends Command final case class GetJobByStatus(status: Status, replyTo: ActorRef[Seq[Job]]) extends Command final case class ClearJobs(replyTo: ActorRef[Response]) extends Command // This behavior handles all possible incoming messages and keeps the state in the function parameter def apply(jobs: Map[Long, Job] = Map.empty): Behavior[Command] = Behaviors.receiveMessage { case AddJob(job, replyTo) if jobs.contains(job.id) => replyTo ! KO("Job already exists") Behaviors.same case AddJob(job, replyTo) => replyTo ! OK JobRepository(jobs.+(job.id -> job)) case GetJobById(id, replyTo) => replyTo ! jobs.get(id) Behaviors.same case ClearJobs(replyTo) => replyTo ! OK JobRepository(Map.empty) } }
Then, let’s define the JSON marshaller and unmarshallers for the HTTP routes:
- Scala
-
source
import akka.http.scaladsl.marshallers.sprayjson.SprayJsonSupport import spray.json.DefaultJsonProtocol import spray.json.DeserializationException import spray.json.JsString import spray.json.JsValue import spray.json.RootJsonFormat trait JsonSupport extends SprayJsonSupport { // import the default encoders for primitive types (Int, String, Lists etc) import DefaultJsonProtocol._ import JobRepository._ implicit object StatusFormat extends RootJsonFormat[Status] { def write(status: Status): JsValue = status match { case Failed => JsString("Failed") case Successful => JsString("Successful") } def read(json: JsValue): Status = json match { case JsString("Failed") => Failed case JsString("Successful") => Successful case _ => throw new DeserializationException("Status unexpected") } } implicit val jobFormat = jsonFormat4(Job) implicit val jobsFormat = jsonFormat1(Jobs) }
Next step is to define the Route
that will communicate with the previously defined behavior and handle all its possible responses:
- Scala
-
source
import akka.actor.typed.ActorSystem import akka.util.Timeout import akka.http.scaladsl.server.Directives._ import akka.http.scaladsl.model.StatusCodes import akka.http.scaladsl.server.Route import scala.concurrent.duration._ import scala.concurrent.Future class JobRoutes(buildJobRepository: ActorRef[JobRepository.Command])(implicit system: ActorSystem[_]) extends JsonSupport { import akka.actor.typed.scaladsl.AskPattern.Askable // asking someone requires a timeout and a scheduler, if the timeout hits without response // the ask is failed with a TimeoutException implicit val timeout: Timeout = 3.seconds // implicit scheduler only needed in 2.5 // in 2.6 having an implicit typed ActorSystem in scope is enough if you import AskPattern.schedulerFromActorSystem implicit val scheduler = system.scheduler lazy val theJobRoutes: Route = pathPrefix("jobs") { concat( pathEnd { concat( post { entity(as[JobRepository.Job]) { job => val operationPerformed: Future[JobRepository.Response] = buildJobRepository.ask(JobRepository.AddJob(job, _)) onSuccess(operationPerformed) { case JobRepository.OK => complete("Job added") case JobRepository.KO(reason) => complete(StatusCodes.InternalServerError -> reason) } } }, delete { val operationPerformed: Future[JobRepository.Response] = buildJobRepository.ask(JobRepository.ClearJobs(_)) onSuccess(operationPerformed) { case JobRepository.OK => complete("Jobs cleared") case JobRepository.KO(reason) => complete(StatusCodes.InternalServerError -> reason) } } ) }, (get & path(LongNumber)) { id => val maybeJob: Future[Option[JobRepository.Job]] = buildJobRepository.ask(JobRepository.GetJobById(id, _)) rejectEmptyResponse { complete(maybeJob) } } ) } }
Finally, we create a Behavior
that bootstraps the web server and use it as the root behavior of our actor system:
- Scala
-
source
import akka.actor.typed.PostStop import akka.actor.typed.scaladsl.adapter._ import akka.stream.ActorMaterializer import akka.http.scaladsl.Http.ServerBinding import akka.http.scaladsl.Http import scala.concurrent.ExecutionContextExecutor import scala.util.{ Success, Failure } object Server { sealed trait Message private final case class StartFailed(cause: Throwable) extends Message private final case class Started(binding: ServerBinding) extends Message case object Stop extends Message def apply(host: String, port: Int): Behavior[Message] = Behaviors.setup { ctx => implicit val system = ctx.system // http doesn't know about akka typed so provide untyped system implicit val untypedSystem: akka.actor.ActorSystem = ctx.system.toClassic // implicit materializer only required in Akka 2.5 // in 2.6 having an implicit classic or typed ActorSystem in scope is enough implicit val materializer: ActorMaterializer = ActorMaterializer()(ctx.system.toClassic) implicit val ec: ExecutionContextExecutor = ctx.system.executionContext val buildJobRepository = ctx.spawn(JobRepository(), "JobRepository") val routes = new JobRoutes(buildJobRepository) val serverBinding: Future[Http.ServerBinding] = Http.apply().bindAndHandle(routes.theJobRoutes, host, port) ctx.pipeToSelf(serverBinding) { case Success(binding) => Started(binding) case Failure(ex) => StartFailed(ex) } def running(binding: ServerBinding): Behavior[Message] = Behaviors.receiveMessagePartial[Message] { case Stop => ctx.log.info( "Stopping server http://{}:{}/", binding.localAddress.getHostString, binding.localAddress.getPort) Behaviors.stopped }.receiveSignal { case (_, PostStop) => binding.unbind() Behaviors.same } def starting(wasStopped: Boolean): Behaviors.Receive[Message] = Behaviors.receiveMessage[Message] { case StartFailed(cause) => throw new RuntimeException("Server failed to start", cause) case Started(binding) => ctx.log.info( "Server online at http://{}:{}/", binding.localAddress.getHostString, binding.localAddress.getPort) if (wasStopped) ctx.self ! Stop running(binding) case Stop => // we got a stop message but haven't completed starting yet, // we cannot stop until starting has completed starting(wasStopped = true) } starting(wasStopped = false) } } def main(args: Array[String]) { val system: ActorSystem[Server.Message] = ActorSystem(Server("localhost", 8080), "BuildJobsServer") }
Note that the akka.actor.typed.ActorSystem
is converted with toClassic
, which comes from import akka.actor.typed.scaladsl.adapter._
. If you are using an earlier version than Akka 2.5.26 this conversion method is named toUntyped
.
Dynamic Routing Example
As the routes are evaluated for each request, it is possible to make changes at runtime. Please note that every access may happen on a separated thread, so any shared mutable state must be thread safe.
The following is an Akka HTTP route definition that allows dynamically adding new or updating mock endpoints with associated request-response pairs at runtime.
- Scala
-
source
case class MockDefinition(path: String, requests: Seq[JsValue], responses: Seq[JsValue]) implicit val format = jsonFormat3(MockDefinition) @volatile var state = Map.empty[String, Map[JsValue, JsValue]] // fixed route to update state val fixedRoute: Route = post { pathSingleSlash { entity(as[MockDefinition]) { mock => val mapping = mock.requests.zip(mock.responses).toMap state = state + (mock.path -> mapping) complete("ok") } } } // dynamic routing based on current state val dynamicRoute: Route = ctx => { val routes = state.map { case (segment, responses) => post { path(segment) { entity(as[JsValue]) { input => complete(responses.get(input)) } } } } concat(routes.toList: _*)(ctx) } val route = fixedRoute ~ dynamicRoute
- Java
For example, let’s say we do a POST request with body:
{
"path": "test",
"requests": [
{"id": 1},
{"id": 2}
],
"responses": [
{"amount": 1000},
{"amount": 2000}
]
}
Subsequent POST request to /test
with body {"id": 1}
will be responded with {"amount": 1000}
.
Handling HTTP Server failures in the High-Level API
There are various situations when failure may occur while initialising or running an Akka HTTP server. Akka by default will log all these failures, however sometimes one may want to react to failures in addition to them just being logged, for example by shutting down the actor system, or notifying some external monitoring end-point explicitly.
Bind failures
For example the server might be unable to bind to the given port. For example when the port is already taken by another application, or if the port is privileged (i.e. only usable by root
). In this case the “binding future” will fail immediately, and we can react to it by listening on the Future
’s completion:
- Scala
-
source
import akka.actor.ActorSystem import akka.http.scaladsl.Http import akka.http.scaladsl.Http.ServerBinding import akka.http.scaladsl.server.Directives._ import akka.stream.ActorMaterializer import scala.concurrent.Future object WebServer { def main(args: Array[String]) { implicit val system = ActorSystem() implicit val materializer = ActorMaterializer() // needed for the future foreach in the end implicit val executionContext = system.dispatcher val handler = get { complete("Hello world!") } // let's say the OS won't allow us to bind to 80. val (host, port) = ("localhost", 80) val bindingFuture: Future[ServerBinding] = Http().bindAndHandle(handler, host, port) bindingFuture.failed.foreach { ex => log.error(ex, "Failed to bind to {}:{}!", host, port) } } }
- Java
For a more low-level overview of the kinds of failures that can happen and also more fine-grained control over them refer to the Handling HTTP Server failures in the Low-Level API documentation.
Failures and exceptions inside the Routing DSL
Exception handling within the Routing DSL is done by providing ExceptionHandler
s which are documented in-depth in the Exception Handling section of the documentation. You can use them to transform exceptions into HttpResponse
s with appropriate error codes and human-readable failure descriptions.
File uploads
For high level directives to handle uploads see the FileUploadDirectives.
Handling a simple file upload from for example a browser form with a file input can be done by accepting a Multipart.FormData entity, note that the body parts are Source rather than all available right away, and so is the individual body part payload so you will need to consume those streams both for the file and for the form fields.
Here is a simple example which just dumps the uploaded file into a temporary file on disk, collects some form fields and saves an entry to a fictive database:
- Scala
-
source
val uploadVideo = path("video") { entity(as[Multipart.FormData]) { formData => // collect all parts of the multipart as it arrives into a map val allPartsF: Future[Map[String, Any]] = formData.parts.mapAsync[(String, Any)](1) { case b: BodyPart if b.name == "file" => // stream into a file as the chunks of it arrives and return a future // file to where it got stored val file = File.createTempFile("upload", "tmp") b.entity.dataBytes.runWith(FileIO.toPath(file.toPath)).map(_ => (b.name -> file)) case b: BodyPart => // collect form field values b.toStrict(2.seconds).map(strict => (b.name -> strict.entity.data.utf8String)) }.runFold(Map.empty[String, Any])((map, tuple) => map + tuple) val done = allPartsF.map { allParts => // You would have some better validation/unmarshalling here db.create(Video( file = allParts("file").asInstanceOf[File], title = allParts("title").asInstanceOf[String], author = allParts("author").asInstanceOf[String])) } // when processing have finished create a response for the user onSuccess(allPartsF) { allParts => complete { "ok!" } } } }
- Java
You can transform the uploaded files as they arrive rather than storing them in a temporary file as in the previous example. In this example we accept any number of .csv
files, parse those into lines and split each line before we send it to an actor for further processing:
- Scala
-
source
val splitLines = Framing.delimiter(ByteString("\n"), 256) val csvUploads = path("metadata" / LongNumber) { id => entity(as[Multipart.FormData]) { formData => val done: Future[Done] = formData.parts.mapAsync(1) { case b: BodyPart if b.filename.exists(_.endsWith(".csv")) => b.entity.dataBytes .via(splitLines) .map(_.utf8String.split(",").toVector) .runForeach(csv => metadataActor ! MetadataActor.Entry(id, csv)) case _ => Future.successful(Done) }.runWith(Sink.ignore) // when processing have finished create a response for the user onSuccess(done) { _ => complete { "ok!" } } } }
- Java
Configuring Server-side HTTPS
For detailed documentation about configuring and using HTTPS on the server-side refer to Server-Side HTTPS Support.