Production-ready template for building asynchronous RabbitMQ workers in Go.

• Message routing via ThreeDotsLabs/watermill + watermill-amqp
• Middleware pipeline: structured logging, configurable retry with backoff, panic recovery, Prometheus metrics
• Diagnostic HTTP server: /ping healthcheck, /metrics (Prometheus), /debug/pprof/* profiling endpoints
• Structured logging via uber-go/zap with optional Sentry integration
• Graceful shutdown on OS signal
• Environment-based configuration via kelseyhightower/envconfig

.
├── cmd/                    # Entrypoint
├── internal/
│   ├── app/
│   │   ├── config/         # Configuration (envconfig)
│   │   ├── logger/         # Zap logger setup
│   │   └── server/         # Diagnostic HTTP server
│   ├── middleware/         # Logging, retry, recoverer, metrics
│   ├── publishers/pub/     # AMQP publisher
│   ├── subscribers/sub/    # AMQP subscriber + message handler
│   ├── marshalers/         # Message marshaling
│   ├── metrics/            # Prometheus metrics
│   ├── httpclients/        # HTTP client wrapper
│   └── retryable/          # Retryable error types
├── scripts/
└── docker-compose.dev.yml

All configuration is done via environment variables:

docker-compose -f docker-compose.dev.yml up -d

make build
make run-server

make run-dev

make test

Or just tests (without lint):

./scripts/test.sh

make golangci

The message handler is located in internal/subscribers/sub/subscriber.go. Implement your business logic inside the Handler method:

func (s *Subscriber) Handler(msg *message.Message) error {
    // your logic here
    return nil
}

To make an error retryable (triggers the retry middleware), wrap it with the retryable error type from internal/retryable/errors.go.

• Branch with cobra + viper CLI — variant using spf13/cobra + spf13/viper

MIT