This channel is available for all Akka enthusiasts—newbies as well as gurus—for the exchange of knowledge and the coordination of efforts around Akka. For more info see: http://bit.ly/akka-gitter
savedSender instead of sender(). Anyway this is just dummy code and my question is more about handling effects: should I use unsafeRunSync() or unsafeToFuture() or is there any better solution? I have library I need to work with and I don't know what are 'best practices' to connect cats-effect with akka actors. With heavy blocking operations solution inside Bar class is desirable but is it acceptable to use unsafeRunSync() with some light computations?
Hello there, I have an idea that I would like to validate with Akka pros.
We have an existing system that manages (let's say) clients and invoices. And we have to integrate that system with other applications. We don't want to pollute our code base with this integration so we have another module that talk with the existing system trough his REST Api (..).
The integration is mainly to:
- execute many commands in one call (CreateInvoice -> CreateClient + CreateInvoice).
- map our identifiers with the external ones.
- translate the domain concepts (Customer = Client + Invoices).
My idea is to create some actors that represents the main concepts, dispatch and coordinate the Http calls to the API to realize the command. (AddCustomer will be GET||POST /clients and one or many POST /invoices).
The thing is that "GET||POST /clients". If a client exist I should create an actor for it and load his state, otherwise I should create a new actor and populate his state. But I cannot have persistent actors in the integration module. I am thinking to use actorSelection() with a fallback to a new actor. Then, for each case, send an "Init" command that is filled from the API result for an existing one or from the incoming command for a new client.
So basically the flow will be:
- HTTP: POST /integration/invoices {clientId, clientName, invoices[]}
- Controller: InvoiceService.getOrCreateClient(clientId, clientName)
2.1. InvoiceService: actorSecelection(clientId).map(_ ! Init(..)).recover( actorOf(..) ! Init(..) ) - Controller: invoices.foreach( clientActor ! AddInvoice(..) )
3.1. ClienActor: WsClient.get("legacy-api/some-check")
3.2.a. ClientActor: on 3.1. match Success => WsClient.post("legacy-api/invoices")
3.2.b. ClientActor: on 3.1. match Failure => WsClient.post("legacy-api/other-resource")
Does that looks very weird or is it a viable approach ?
Thanks
map and I can't find the right syntax. More info https://travis-ci.com/akka/alpakka/jobs/226125470#L850
How can I set positive-ttl in
AsyncDnsResolverlike inInetAddressDnsResolver?
I've faced with issue when a dns-server response contains ttl, for instance, a 1 day, but ip of the target service could be changed in any time.InetAddressDnsResolverrefreshes cache and sends requests to the dns-server according to the positive-ttl or negative-ttl settings. ButAsyncDnsResolverdon't send any request until cache record expired, in my case, it's a 1 day. I've triedInetAddressDnsResolverwithDnsServiceDiscoverybutDnsServiceDiscoveryusesDnsProtocol.ResolveandInetAddressDnsResolvercan process onlyDns.Resolve.
Is there a way to deal with this issue?
Can anyone help with this?
DnsResolverwith DnsProvider which will be able to handle DnsProtocol.Resolveand positive-ttlin the same time . Unfortunately, most of all components, like DnsClient, belong to InternalApi, and I have to implement this components by myself.
@volgad Also definitely read https://doc.akka.io/docs/akka/current/stream/stream-rate.html and https://doc.akka.io/docs/akka/current/stream/stream-parallelism.html
After all this you can also consider and try to increase the input buffer sizes for Akka Streams stages, see https://doc.akka.io/docs/akka/current/stream/stream-rate.html#internal-buffers-and-their-effect (part of the second link above).
I am implenting a 'message queue like system' using Akka actor router. https://doc.akka.io/docs/akka/current/routing.html
I want a routing policy that sends only one message to a idle child actor. (If it's already processing a message, then I would like to send a message to a different child actor. If all the children actors are busy, then the router stashes messages until a child actor is available)
Is there a routing policy for such need? Thank you :)
@cosmir17 there's no way for the router to know what the children are doing, but you might want to look at the work pulling pattern where children request the next message when they're idle: https://www.michaelpollmeier.com/akka-work-pulling-pattern
You might also be able to solve your problems using Akka Streams.
ActorRef source type https://doc.akka.io/docs/akka/current/stream/operators/Source/actorRef.html
Hi, I have a question.
I have an Akka class
class GenericActorClass(headphone: String) extends Actor {
....
}
val actorRef = system.actorOf(props.....)how do I get a value of 'headphone' variable from actorRef?
Thank you,
Sean
Hi, I am making http-requests(async-http-client) with different paths inside my Actor class.
def recieve: Receive = {
case .... =>
val sdr = sender()
httpClient.get(path).params(paramMap).execute().pipeTo(self)(sdr)
}This returns a RichResponse, I can pattern matching inside recieve method. I can get the full path of the URL from RichResponse. Or if it doesn't, I can map Future and include path as a tuple.
when a request times out, it becomes an issue because I can't know which request times out (different paths) and there seems no way to pass the full path URL.
case r@ Status.Failure(_: TimeoutException) =>the timeout exception only shows 'www.domain.com', it doesn't show a full path.
Is there a way to pass 'a full url path' inside Akka.Status.Failure?
@jacobg by default, Akka will trigger CoordinatedShutdown in the JVM shutdown hook, which will be run when it gets SIGTERM. This can be disabled with the akka.coordinated-shutdown.run-by-jvm-shutdown-hook=off config setting. See https://doc.akka.io/docs/akka/current/actors.html#coordinated-shutdown for details.
However, Akka HTTP does not register a CoordinatedShutdown task by default, so you should do that yourself. See akka/akka-http#1210 for discussion.
The thing is that the WS may reply a 404 in which case I should send another request to save teh values on the WS.
keep-alive? If yes, how to specify it? I cannot find it anywhere in https://doc.akka.io/docs/akka-http/current/introduction.html. It mentioned keep-alive for websocket only.