A distributed system monitoring solution using Akka.NET Remoting + Streams, ASP.NET Core 8, SignalR, and Angular with Tailwind CSS.

┌─────────────────────────────────────────────────────────────────────────┐ │ Architecture Overview │ ├─────────────────┐ Akka Remote TCP ┌────────────────────────────┐ │ │ Machine A │ ──────────────────► │ MetricsHub │ │ │ Collector │ PushMetrics msg │ (Central Akka Actor) │ │ │ (Akka Actor + │ ◄────────────────── │ Receives from all nodes │ │ │ Streams tick) │ RegisterCollector │ ──────────────────────► │ │ │ REST API │ │ SignalR push to UI │ │ └─────────────────┘ └────────────────────────────┘ │ │ │ ┌─────────────────┐ Akka Remote TCP │ WebSocket │ │ Machine B │ ──────────────────► ▼ │ │ Collector │ ┌────────────────────────────┐ │ │ │ │ Angular UI │ │ │ REST API │ ◄───── HTTP ──────── │ SignalR Client │ │ │ (start/stop) │ │ Tailwind dashboard │ │ └─────────────────┘ └────────────────────────────┘ │ └─────────────────────────────────────────────────────────────────────────┘ 

# From the repo root docker-compose up --build

• UI: http://localhost
• Hub API + Swagger: http://localhost:5000/swagger
• Collector 1 API: http://localhost:5001
• Collector 2 API: http://localhost:5002

Start collection on node 1:

curl -X POST http://localhost:5001/api/metrics/start -H "Content-Type: application/json" -d '{"intervalMs": 2000}'

1. Start MetricsHub

cd MetricsHub dotnet run # Listening on http://localhost:5000 # Akka TCP on port 9080

2. Start a Collector (on any machine)

cd MetricsCollector # Edit appsettings.json: set HubAddress to the hub's address dotnet run # REST API on http://localhost:5001

3. Start the Angular UI

cd angular-ui npm install npm start # Open http://localhost:4200

4. Start collecting metrics via REST API

# Start collection (default: every 2 seconds) curl -X POST http://localhost:5001/api/metrics/start # With custom interval curl -X POST http://localhost:5001/api/metrics/start -H "Content-Type: application/json" -d '{"intervalMs": 1000}' # Get status curl http://localhost:5001/api/metrics/status # Stop collection curl -X POST http://localhost:5001/api/metrics/stop

1. Build the collector: dotnet publish MetricsCollector -c Release
2. Copy to target machine
3. Set environment variables or edit appsettings.json:

   { "Akka": { "NodeId": "machine-name-or-uuid", "Port": "8081", "PublicHostname": "192.168.1.X", "HubAddress": "akka.tcp://MetricsHub@<HUB_IP>:9080/user/metrics-hub" } }

4. Run: dotnet MetricsCollector.dll
5. Start collection: POST http://<collector-ip>:5001/api/metrics/start

Environment variable override example:

Akka__NodeId=prod-server-1 \ Akka__PublicHostname=10.0.1.50 \ Akka__HubAddress="akka.tcp://MetricsHub@10.0.1.100:9080/user/metrics-hub" \ dotnet MetricsCollector.dll

• MetricsHub starts an ActorSystem named MetricsHub listening on TCP port 9080
• Each MetricsCollector creates an ActorSystem named MetricsCollector
• Collectors resolve the hub actor via: akka.tcp://MetricsHub@<host>:9080/user/metrics-hub
• Messages are serialized with NewtonSoftJsonSerializer

Source.Tick(interval, interval, "tick") .ViaMaterialized(KillSwitches.Single<string>(), Keep.Right) .Select(_ => new CollectTick()) .ToMaterialized(Sink.ActorRef<CollectTick>(self, new StreamCompleted()), Keep.Both) .Run(materializer);

The stream ticks at the configured interval, sends messages to the actor itself, which then collects metrics and pushes to the hub.

POST /api/metrics/start body:

{ "intervalMs": 2000 }

{ "Akka": { "Port": "9080", "Hostname": "0.0.0.0", "PublicHostname": "localhost" }, "Kestrel": { "Endpoints": { "Http": { "Url": "http://0.0.0.0:5000" } } } }

{ "Akka": { "NodeId": "node-1", "Port": "8081", "PublicHostname": "localhost", "HubAddress": "akka.tcp://MetricsHub@localhost:9080/user/metrics-hub" }, "Kestrel": { "Endpoints": { "Http": { "Url": "http://0.0.0.0:5001" } } } }

Collectors need outbound access to hub port 9080. Hub needs inbound access on port 9080.