sup

sup

sup is a high-performance, low-allocation Actor Model library for Go.

It provides a robust foundation for building highly concurrent, distributed, and fault-tolerant stateful applications. It achieves zero-allocation for asynchronous messages (Cast) and minimizes overhead for synchronous requests (Call) by utilizing internal resource pooling. It embraces standard Go idioms (select, channels, and context) rather than hiding them behind heavy frameworks.

Features

• Idiomatic Go — Actors are just goroutines running a select loop. No magic interfaces, no reflection, no global registries.
• OTP-style supervision — Erlang-inspired supervisor trees with Permanent, Transient, and Temporary restart policies.
• Panic recovery — Panics are caught, wrapped with a stack trace, and reported via WithOnError. The actor is then restarted according to the policy.
• Restart limits — Optionally cap restarts within a sliding time window with WithRestartLimit.
• No goroutine leaks — context.Context integration ensures all actors shut down cleanly when the parent context is canceled.
• Supervisor trees — Supervisors implement the Actor interface, so they can be nested inside other supervisors.

Installation

go get github.com/webermarci/sup

Quick Start

package main

import (
	"context"
	"fmt"
	"time"

	"github.com/webermarci/sup"
)

// 1. Define internal messages
type incrementMsg struct{ amount int }
type getCountMsg struct{}

// 2. Define your Actor
type Counter struct {
	*sup.BaseActor
	mailbox *sup.Mailbox
	count int
}

func NewCounter() *Counter {
	return &Counter{
		BaseActor: sup.NewBaseActor("counter"),
		mailbox: sup.NewMailbox(10),
	}
}

// 3. Clean public API — callers never interact with the mailbox directly
func (c *Counter) Increment(amount int) {
	_ = sup.Cast(c.Mailbox, incrementMsg{amount: amount})
}

func (c *Counter) Get() (int, error) {
	return sup.Call[getCountMsg, int](c.Mailbox, getCountMsg{})
}

// 4. The Run loop is a standard Go select statement
func (c *Counter) Run(ctx context.Context) error {
	for {
		select {
		case <-ctx.Done():
			return ctx.Err()
		case msg := <-c.mailbox.Receive():
			switch m := msg.(type) {
			case sup.CastRequest[incrementMsg]:
				c.count += m.Payload().amount
			case sup.CallRequest[getCountMsg, int]:
				m.Reply(c.count, nil)
			}
		}
	}
}

func main() {
	ctx, cancel := context.WithCancel(context.Background())
	defer cancel()

	counter := NewCounter()

	supervisor := sup.NewSupervisor("root",
		sup.WithActor(counter),
		sup.WithPolicy(sup.Permanent),
		sup.WithRestartDelay(time.Second),
		sup.WithRestartLimit(5, 10*time.Second),
		sup.WithOnError(func(actor sup.Actor, err error) {
			fmt.Printf("Actor %s failed with error: %v\n", actor.Name(), err)
		}),
	)

	go supervisor.Run(ctx)

	counter.Increment(10)
	counter.Increment(32)

	count, err := counter.Get()
	if err != nil {
		panic(err)
	}

	fmt.Printf("Final count: %d\n", count) // Final count: 42

	cancel()
	supervisor.Wait()
}

Restart Policies

Mailbox

A Mailbox is the actor's message queue. Messages are sent using Cast (fire-and-forget) or Call (request-reply), and received inside the actor's Run loop via Receive().

Sending variants

Supervisor Trees

Supervisors implement the Actor interface, so they compose naturally into trees. When the root context is canceled, shutdown propagates recursively through the entire tree.

dbActor := NewDatabaseActor()
cacheActor := NewCacheActor()

// Child supervisor manages data-layer actors
dataSup := sup.NewSupervisor("data_supervisor",
	sup.WithActors(dbActor, cacheActor),
	sup.WithPolicy(sup.Permanent),
	sup.WithRestartDelay(500*time.Millisecond),
)

// Root supervisor treats the child supervisor as an actor
root := sup.NewSupervisor("root",
	sup.WithActor(dataSup),
	sup.WithPolicy(sup.Permanent),
)

go root.Run(ctx)

Stateless Actors

For actors that don't need a mailbox or internal state, use ActorFunc:

healthCheck := sup.ActorFunc("health", func(ctx context.Context) error {
	ticker := time.NewTicker(30 * time.Second)
	defer ticker.Stop()
	for {
		select {
		case <-ctx.Done():
			return ctx.Err()
		case <-ticker.C:
			if err := ping(); err != nil {
				return err // supervisor will restart based on policy
			}
		}
	}
})

sup.NewSupervisor("health_supervisor",
	sup.WithActor(healthCheck),
	sup.WithPolicy(sup.Transient),
).Run(ctx)

Dynamic Spawning

Use Spawn to start actors dynamically after the supervisor is already running:

supervisor := sup.NewSupervisor("job_supervisor",
	sup.WithPolicy(sup.Temporary),
)
go supervisor.Run(ctx)

// Later, spawn actors on demand
for _, job := range jobs {
	supervisor.Spawn(ctx, newJobActor(job))
}

supervisor.Wait()

Packages

• sup — Core supervisor and mailbox implementation
• sup/bus — Higher-level abstractions for polling and controlling with automatic scheduling and change notifications
• sup/exec — Actor that manages external processes with graceful shutdown and output redirection

Benchmark

goos: darwin
goarch: arm64
pkg: github.com/webermarci/sup
cpu: Apple M5
Benchmark_Cast-10                       20725142    57.4 ns/op     0 B/op   0 allocs/op
Benchmark_Cast_Concurrent-10            10265941   117.4 ns/op     0 B/op   0 allocs/op
Benchmark_CastContext-10                21805774    55.2 ns/op     0 B/op   0 allocs/op
Benchmark_CastContext_Concurrent-10     15253065    79.0 ns/op     0 B/op   0 allocs/op
Benchmark_CastContext_Expired-10        23459665    51.0 ns/op     0 B/op   0 allocs/op
Benchmark_TryCast-10                   198711219     6.0 ns/op     0 B/op   0 allocs/op
Benchmark_TryCast_Concurrent-10         89085205    13.9 ns/op     0 B/op   0 allocs/op
Benchmark_TryCast_Full-10              254923090     4.7 ns/op     0 B/op   0 allocs/op
Benchmark_Call-10                        3095778   390.2 ns/op   152 B/op   3 allocs/op
Benchmark_Call_Concurrent-10             2198695   506.8 ns/op   152 B/op   3 allocs/op
Benchmark_CallContext-10                 2545513   451.1 ns/op   152 B/op   3 allocs/op
Benchmark_CallContext_Concurrent-10      1301760   896.5 ns/op   152 B/op   3 allocs/op
Benchmark_CallContext_Expired-10        14855138    78.4 ns/op   152 B/op   3 allocs/op
Benchmark_TryCall-10                     3060210   392.7 ns/op   152 B/op   3 allocs/op
Benchmark_TryCall_Concurrent-10          2227587   527.4 ns/op   152 B/op   3 allocs/op
Benchmark_Supervisor_SpawnAndExit-10     4714188   236.9 ns/op    72 B/op   2 allocs/op