NATS Logo by Example

Pull Consumers in JetStream

A pull consumer allows for the application to fetch one or more messages on-demand using a subscription bound to the consumer. This allows the application to control the flow of the messages coming in so it can process and ack them in an appropriate amount of time.

A consumer can either be durable or ephemeral. A durable consumer will have its state tracked on the server, most importantly, the last acknowledged message from the client.

Ephemeral consumers are useful as one-off needs and are a bit cheaper in terms of resources and management. However, ephemerals do not (of course) persist after the primary subscriber unsubscribes. The server will automatically clean up (delete) the consumer after a period of time.

Since each subscription is fetching messages on-demand, multiple subscriptions can be create bound to the same pull consumer without any additional configuration. Each subscriber can fetch batches of messages and process them concurrently.

It is important to note that the messages in a given batch are ordered with respect to each other, but each subscriber will be handling a batch independently. If there is a need to have determinstic partitioning for scalable order processing, learn more here.

CLI Go Python JavaScript Rust C# C#2 Java Ruby Elixir Crystal C
Jump to the output or the recording
$ nbe run jetstream/pull-consumer/go
View the source code or learn how to run this example yourself


package main

import (


func main() {

Use the env variable if running in the container, otherwise use the default.

	url := os.Getenv("NATS_URL")
	if url == "" {
		url = nats.DefaultURL

Create an unauthenticated connection to NATS.

	nc, _ := nats.Connect(url)
	defer nc.Drain()

Access JetStream for managing streams and consumers as well as for publishing and consuming messages to and from the stream.

	js, _ := jetstream.New(nc)

	streamName := "EVENTS"

JetStream API uses context for timeouts and cancellation.

	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
	defer cancel()

Declare a simple limits-based stream.

	stream, _ := js.CreateStream(ctx, jetstream.StreamConfig{
		Name:     streamName,
		Subjects: []string{"events.>"},

Publish a few messages for the example.

	js.Publish(ctx, "events.1", nil)
	js.Publish(ctx, "events.2", nil)
	js.Publish(ctx, "events.3", nil)

Create the consumer bound to the previously created stream. If durable name is not supplied, consumer will be removed after InactiveThreshold (defaults to 5 seconds) is reached when not actively consuming messages. Name is optional, if not provided it will be auto-generated. For this example, let’s use the consumer with no options, which will be ephemeral with auto-generated name.

	cons, _ := stream.CreateOrUpdateConsumer(ctx, jetstream.ConsumerConfig{})

Messages can be consumed continuously in callback using Consume method. Consume can be supplied with various options, but for this example we will use the default ones. WaitGroup is used as part of this example to make sure to stop processing after we process 3 messages (so that it does not interfere with other examples).

	wg := sync.WaitGroup{}

	cc, _ := cons.Consume(func(msg jetstream.Msg) {
		fmt.Println("received msg on", msg.Subject())

Consume can be stopped by calling Stop on the returned ConsumerContext. This will stop the callback from being called and stop retrieving the messages.


Publish more messages.

	js.Publish(ctx, "events.1", nil)
	js.Publish(ctx, "events.2", nil)
	js.Publish(ctx, "events.3", nil)

We can fetch messages in batches. The first argument being the batch size which is the maximum number of messages that should be returned. For this first fetch, we ask for two and we will get those since they are in the stream.

	msgs, _ := cons.Fetch(2)
	var i int
	for msg := range msgs.Messages() {

Let’s ack the messages so they are not redelivered.

	fmt.Printf("got %d messages\n", i)

Fetch puts messages on the returned Messages() channel. This channel will only be closed when the requested number of messages have been received or the operation times out. If we do not want to wait for the rest of the messages and want to quickly return as many messages as there are available (up to provided batch size), we can use FetchNoWait instead. Here, because we have already received two messages, we will only get one more.

	msgs, _ = cons.FetchNoWait(100)
	i = 0
	for msg := range msgs.Messages() {
	fmt.Printf("got %d messages\n", i)

Finally, if we are at the end of the stream and we call fetch, the call will be blocked until the “max wait” time which is 5 seconds by default, but this can be set explicitly as an option.

	fetchStart := time.Now()
	msgs, _ = cons.Fetch(1, jetstream.FetchMaxWait(time.Second))
	i = 0
	for msg := range msgs.Messages() {

	fmt.Printf("got %d messages in %v\n", i, time.Since(fetchStart))

Durable consumers can be created by specifying the Durable name. Durable consumers are not removed automatically regardless of the InactiveThreshold. They can be removed by calling DeleteConsumer.

	dur, _ := stream.CreateOrUpdateConsumer(ctx, jetstream.ConsumerConfig{
		Durable: "processor",

Consume and fetch work the same way for durable consumers.

	msgs, _ = dur.Fetch(1)
	msg := <-msgs.Messages()
	fmt.Printf("received %q from durable consumer\n", msg.Subject())

While ephemeral consumers will be removed after InactiveThreshold, durable consumers have to be removed explicitly if no longer needed.

	stream.DeleteConsumer(ctx, "processor")

Let’s try to get the consumer to make sure it’s gone.

	_, err := stream.Consumer(ctx, "processor")

	fmt.Println("consumer deleted:", errors.Is(err, jetstream.ErrConsumerNotFound))


received msg on events.1
received msg on events.2
received msg on events.3
got 2 messages
got 1 messages
got 0 messages in 1.001453292s
received "events.1" from durable consumer
consumer deleted: true


Note, playback is half speed to make it a bit easier to follow.