Queue

Queue provides durable message queuing for asynchronous processing. You send messages to a queue, and they get processed by consumer functions at a later time.

Queues work with Observers to handle message processing. Only a single observer per queue is allowed. When messages arrive, your observer function receives them for processing.

Messages support multiple content types including text, JSON, binary data, and V8-serialized objects. The system handles message persistence, delivery guarantees, and automatic retries.

Creating

Define queues in your raindrop.manifest:

application "demo-app" {
  queue "demo-queue" {}
}

Accessing

Access your queue through environment bindings in your code:

export default {
  async fetch(request, env) {
    await env.DEMO_QUEUE.send({ userId: 123, action: "process" });
    return new Response("Message sent");
  }
};

Core Concepts

Main Interfaces

Queue - Send individual messages or batches to the queue
Message - Represents a received message with processing controls
MessageBatch - Group of messages processed together

Core Data Types

QueueContentType

// Supported message content formats
type QueueContentType = 'text' | 'bytes' | 'json' | 'v8';

MessageSendRequest

// Configuration for sending a message in a batch
interface MessageSendRequest<Body = unknown> {
  body: Body;                    // Message content
  contentType?: QueueContentType; // Content format
  delaySeconds?: number;         // Processing delay
}

QueueSendOptions

// Options for sending individual messages
interface QueueSendOptions {
  contentType?: QueueContentType; // Message format
  delaySeconds?: number;         // Processing delay (default: 0)
}

QueueSendBatchOptions

// Options for batch message sending
interface QueueSendBatchOptions {
  delaySeconds?: number; // Delay for all messages in batch
}

QueueRetryOptions

// Configuration for message retry attempts
interface QueueRetryOptions {
  delaySeconds?: number; // Delay before retry
}

System Limits

Maximum message size: 128KB
Maximum consumer batch size: 100 messages
Maximum messages per sendBatch call: 100 (or 256KB total)
Maximum message retries: 100
Maximum batch wait time: 60 seconds
Per-queue throughput: 5,000 messages per second
Message retention period: 4 days (default)

send(message: Body, options?: QueueSendOptions): Promise<void>

Promise<void> // Resolves when message is queued

Example

Send a JSON message to the queue:

// Send user registration data for processing
await env.USER_QUEUE.send(
  { userId: 456, email: "user@example.com", action: "welcome" },
  { contentType: "json", delaySeconds: 30 }
);

sendBatch

Input
Output

sendBatch(
  messages: Iterable<MessageSendRequest<Body>>,
  options?: QueueSendBatchOptions
): Promise<void>

Promise<void> // Resolves when all messages are queued

Example

Send multiple messages with different configurations:

// Send batch of notification messages
const messages = [
  { body: { type: "email", recipient: "user1@example.com" }, contentType: "json" },
  { body: { type: "sms", recipient: "+1234567890" }, delaySeconds: 60 },
  { body: "Push notification content", contentType: "text" }
];

await env.NOTIFICATION_QUEUE.sendBatch(messages, { delaySeconds: 10 });

ack

Input
Output

ack(): void

void // Marks message as successfully processed

Example

Acknowledge successful message processing:

// Process message and acknowledge completion
export default {
  async queue(batch, env) {
    for (const message of batch.messages) {
      const data = message.body;
      await processUserData(data);
      message.ack(); // Mark as successfully processed
    }
  }
};

retry

Input
Output

retry(options?: QueueRetryOptions): void

void // Schedules message for retry processing

Example

Retry message processing with delay:

// Retry failed message processing after 120 seconds
export default {
  async queue(batch, env) {
    for (const message of batch.messages) {
      try {
        await processMessage(message.body);
        message.ack();
      } catch (error) {
        message.retry({ delaySeconds: 120 });
      }
    }
  }
};

ackAll

Input
Output

ackAll(): void

void // Marks all messages in batch as processed

Example

Acknowledge all messages in a batch:

// Process entire batch and acknowledge all messages
export default {
  async queue(batch, env) {
    const results = await Promise.allSettled(
      batch.messages.map(msg => processMessage(msg.body))
    );

    if (results.every(r => r.status === "fulfilled")) {
      batch.ackAll(); // All messages processed successfully
    }
  }
};

retryAll

Input
Output

retryAll(options?: QueueRetryOptions): void

void // Schedules all messages for retry processing

Example

Retry all messages in a batch:

// Retry entire batch if processing fails
export default {
  async queue(batch, env) {
    try {
      await processBatch(batch.messages);
      batch.ackAll();
    } catch (error) {
      batch.retryAll({ delaySeconds: 300 }); // Retry in 5 minutes
    }
  }
};