Akka discovery

Skill vezril/claude-toolkit/skills/akka-discovery

Akka Discovery (Akka Core 2.10.x, akka-discovery) in Scala and Java — a pluggable service-discovery SPI that decouples endpoint lookup from configuration, with built-in DNS, config, and aggregate methods (and Kubernetes/AWS/Consul via Akka Management). Use whenever resolving service endpoints at runtime, configuring how an app finds other services across environments, setting up Akka Cluster Bootstrap (automatic seed-node discovery), or pointing an Akka gRPC / Alpakka Kafka client at a discovered service — even if "discovery" isn't named but service location, DNS SRV lookup, or cluster bootstrap are involved.From its SKILL.md

Install

npx -y skills add vezril/claude-toolkit --skill akka-discovery

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.3 KB, as published. Nobody here has run it

Akka Discovery

A small, pluggable service-discovery SPI (akka-discovery) that decouples how you find a service's endpoints from your code, so the same code resolves differently per environment. Built-in methods: config (static HOCON), DNS (A/AAAA + SRV), and aggregate (try several in order). [[akka]] Management adds Kubernetes, AWS EC2/ECS, Consul, and Marathon implementations. It underpins Cluster Bootstrap ([[akka-cluster]] formation without hardcoded seed-nodes) and is used by [[akka-grpc]] and [[alpakka]] Kafka clients.

Dependency: "com.typesafe.akka" %% "akka-discovery" % AkkaVersion. Cross-links: [[akka]] (meta), [[akka-cluster]], [[akka-grpc]], [[alpakka]].

API

Load the Discovery extension to get a ServiceDiscovery, then lookup:

import akka.discovery.{Discovery, Lookup, ServiceDiscovery}
import scala.concurrent.duration._
val sd: ServiceDiscovery = Discovery(system).discovery
val resolved: Future[ServiceDiscovery.Resolved] =
  sd.lookup(Lookup("my-service").withPortName("remoting").withProtocol("tcp"), resolveTimeout = 3.seconds)
// or just: sd.lookup("my-service", 1.second)
// resolved.serviceName, resolved.addresses -> Seq[ResolvedTarget(host, port, address)]

ServiceDiscovery sd = Discovery.get(system).discovery();
CompletionStage<ServiceDiscovery.Resolved> resolved =
  sd.lookup(Lookup.create("my-service").withPortName("remoting").withProtocol("tcp"), Duration.ofSeconds(3));

A Lookup has a mandatory serviceName and optional portName/protocol (used when one service exposes multiple ports). A Resolved holds a list of ResolvedTarget(host, port, address).

Methods (HOCON)

# DNS — async-dns resolver; all three Lookup fields set => SRV query `_portName._protocol.serviceName`,
# otherwise an A/AAAA query on serviceName.
akka.discovery.method = akka-dns

# Config — static endpoints; ignores all Lookup fields but serviceName.
akka.discovery {
  method = config
  config.services {
    service1.endpoints = [ { host = "host1", port = 1233 }, { host = "host2", port = 1234 } ]
  }
}

# Aggregate — try methods in order, falling back when one yields nothing.
akka.discovery {
  method = aggregate
  aggregate.discovery-methods = ["akka-dns", "config"]
}

When to use it

Cluster Bootstrap — instead of hardcoding akka.cluster.seed-nodes, let Cluster Bootstrap (Akka Management) use Discovery (Kubernetes API, DNS SRV, etc.) to find the initial contact points and form the cluster. The main reason this module matters in production.
Client endpoint resolution — point an [[akka-grpc]] client (GrpcClientSettings.usingServiceDiscovery(...)) or [[alpakka]] Kafka (bootstrap.servers via discovery) at a service by name.
Environment portability — start with config in dev, switch to kubernetes-api/aws-api in prod by config alone, no code change.

Gotchas

A discovery method config entry must point to a config location under akka.discovery with at least a class (FQCN) property (since Akka Management 1.0.0; older bare-class-name form is incompatible).
DNS discovery always uses Akka's native async-dns resolver regardless of akka.io.dns.resolver.
SRV terminology mapping: SRV service = your portName, SRV protocol = protocol, SRV name = serviceName; Akka adds the _ prefixes. SRV weights are currently ignored.

[[akka]] — meta skill and module map.
[[akka-cluster]] — Cluster Bootstrap uses Discovery to form a cluster automatically.
[[akka-grpc]], [[alpakka]] — clients can resolve services via Discovery.
Source: Akka Core docs, https://doc.akka.io/libraries/akka-core/current/discovery/index.html (v2.10.x).