@blokcert/node-red-contrib-delayed-job

A Node-RED node for Redis-backed delayed job queues using BullMQ.

Each node acts as both producer and consumer — messages are serialized into a Redis queue on input, and a background worker picks them up and sends them to the output.

Features

• Delayed execution — schedule jobs to run after a specified delay
• Job priority — lower number = higher priority
• TTL (Time-to-live) — automatically discard expired jobs
• Rate limiting — control the interval between job processing
• Concurrency — process multiple jobs simultaneously
• Crash recovery — jobs persist in Redis; restarting Node-RED picks up unfinished jobs
• Distributed workers — multiple Node-RED instances sharing the same queue name form a worker pool
• Buffer support — Buffer objects are preserved through serialization
• i18n — English and Traditional Chinese (zh-TW)

Install

cd ~/.node-red
npm install @blokcert/node-red-contrib-delayed-job

Requires a running Redis server (or compatible, e.g. Valkey, Dragonfly).

Nodes

delayed-job

The main node. Drag it into your flow, configure the Redis connection, and wire it up.

Input: messages are serialized and added to the Redis queue.

Output: when a job is picked up by the worker, the deserialized message is sent out.

Node Settings

| Setting | Default | Description | |---|---|---| | Redis | (required) | Redis connection config | | Concurrency | 1 | Number of jobs processed simultaneously | | Interval | 0 ms | Minimum time between job processing (rate limit) |

Message Properties

| Property | Type | Description | |---|---|---| | msg.delay | number | Delay in milliseconds before the job is processed | | msg.priority | number | Job priority (lower = higher priority) | | msg.ttl | number | Time-to-live in seconds. Expired jobs are discarded. Default: 7 days | | msg.keepAfter | number | Keep job in Redis for N seconds after completion. Default: 7 days |

delayed-job-redis (Config Node)

Shared Redis connection configuration.

| Setting | Default | Description | |---|---|---| | Host | 127.0.0.1 | Redis server hostname | | Port | 6379 | Redis server port | | Password | — | Authentication password (stored encrypted) | | DB | 0 | Redis database index | | TLS | off | Enable TLS/SSL | | Queue Name | default | Queue name prefix. Same name = shared queue across instances |

How It Works

[inject] → [delayed-job] → [debug]
              ↕ Redis

1. A message arrives at the input
2. The node serializes it (removing _msgid) and adds it to the BullMQ queue in Redis
3. The background worker picks up the job, deserializes the message, and sends it to the output
4. Node-RED assigns a new _msgid

Queue Isolation

Each node instance gets a unique queue: {queueName}-{nodeId}. To share a queue across multiple Node-RED instances (distributed worker pool), use the same Queue Name in the Redis config.

Crash Recovery

Jobs are stored in Redis and retained for 7 days after completion by default. If Node-RED crashes or restarts, the worker will pick up any remaining jobs automatically.

Limitations

• msg.req and msg.res (HTTP request/response objects) cannot be serialized and will be lost
• Non-JSON-serializable properties will be dropped

Example Flow

[
  {
    "id": "inject1",
    "type": "inject",
    "payload": "hello",
    "wires": [["delayed1"]]
  },
  {
    "id": "delayed1",
    "type": "delayed-job",
    "name": "my queue",
    "wires": [["debug1"]]
  },
  {
    "id": "debug1",
    "type": "debug",
    "name": "output"
  }
]

Testing

Requires a local Redis server running.

npm test

Set REDIS_HOST and REDIS_PORT environment variables to use a non-default Redis server.

License

MIT