Telegram Bot Service

A consumer application that listens to notification queues and delivers push notifications to users via Telegram

Telegram Bot Service

The Telegram Bot Service functions as a message consumer that processes notifications from RabbitMQ and routes them to subscribed Telegram users. The system handles formatting conversions, maintains subscription caches, and implements automatic retry mechanisms for failed deliveries.

Core Responsibilities

The service manages three primary functions:

• Consuming notifications from message queues
• Converting notification content into Telegram-compatible HTML format
• Distributing messages to all users subscribed to each account

It additionally caches subscription data to minimize database queries.

How to Run

Development

# Start RabbitMQ and PostgreSQL containers
docker-compose up -d queue db

# Set the Telegram bot token
export TELEGRAM_SECRET=your_bot_token

# Start the Telegram service
yarn telegram

Production

# Build the application
yarn build telegram

# Start the service
TELEGRAM_SECRET=your_bot_token node dist/apps/telegram/main.js

Docker

# Build the Docker image
docker build -f apps/telegram/Dockerfile . -t telegram

# Run the container
docker run \
  -e TELEGRAM_SECRET=your_bot_token \
  -e QUEUE_HOST=rabbitmq \
  -e DATABASE_HOST=postgres \
  telegram

Configuration

Setting	Purpose	Mandatory
TELEGRAM_SECRET	Bot authentication token	Yes
QUEUE_HOST	RabbitMQ hostname	No
DATABASE_HOST	PostgreSQL location	No

Obtain bot tokens by messaging @BotFather on Telegram and running /newbot.

Operational Architecture

The service operates through an infinite loop that establishes RabbitMQ connections and processes incoming notifications. When connection losses occur, automatic reconnection attempts trigger after 10-second intervals.

Subscription caching reduces database strain: subscription data refreshes every 5 minutes, with lookups using account addresses as cache keys.

Message Transformation Process

Notifications undergo multi-step conversion:

1. Markdown formatting
2. HTML parsing via the marked library
3. Sanitization with DOMPurify (preserving only Telegram-supported tags)
4. Delivery via Telegram Bot API

Permitted HTML elements include <b>, <i>, <u>, <s>, <a>, <code>, and <pre> tags.

Delivery Workflow

1. Blockchain events trigger notification generation
2. Notification enters the RabbitMQ queue
3. The Telegram service retrieves subscribed chat IDs
4. Content is formatted as HTML
5. Messages are dispatched
6. Successful processing is acknowledged back to the queue

Error Management

• Connection interruptions trigger automatic reconnection cycles
• Message delivery errors are logged without crashing the service
• Failed sanitization attempts include source markdown and problematic HTML in error reports

Subscription Operations

Subscriptions map Ethereum addresses to Telegram chat IDs. Users authenticate via wallet before providing chat IDs through the API service (not through the bot directly). The system automatically incorporates new subscriptions after cache expiration.

Monitoring Output

Structured logging includes notifications about:

• Consumer initialization
• Subscription lookups
• Message dispatching
• Connection status changes

Across debug, info, warning, and error severity levels.