Skip to content

babylonlabs-io/staking-api-service

Folders and files

NameName
Last commit message
Last commit date

Latest commit

c5a2136 · Nov 29, 2024
Nov 29, 2024
Oct 23, 2024
Nov 19, 2024
Nov 29, 2024
Nov 19, 2024
Nov 29, 2024
Apr 17, 2024
Nov 29, 2024
Nov 29, 2024
Mar 20, 2024
Mar 18, 2024
Sep 3, 2024
Aug 16, 2024
Oct 22, 2024
Aug 1, 2024
Sep 3, 2024
Oct 23, 2024
Nov 29, 2024
Nov 29, 2024

Repository files navigation

Staking API Service

The Staking API Service is a critical component of the Babylon Phase-1 system, focused on serving information about the state of the network and receiving unbonding requests for further processing. The API can be utilised by user facing applications, such as staking dApps.

Architectural Design

Staking Service Architecture

This architecture is centered around a message-driven approach, utilizing RabbitMQ queues for inter-service communication. Such a design facilitates high concurrency and enhances fault tolerance by allowing for horizontal scaling and leveraging RabbitMQ's message retry features. The primary infrastructure components include:

  1. MongoDB
  2. RabbitMQ
  3. Redis cache (Work In Progress)

Infra recommendation

Since the service is public-facing and exposes a set of endpoints, it's crucial to apply additional protection mechanisms in front of the service. Here are some recommendations beyond the basic configurations of request size limit, header limit, request caching, and DDOS protection:

  • Request size limit
  • Header limit
  • Request caching
  • Rate limiting
  • DDOS protection mechanism
  • SSL/TLS
  • ... other common API gateway setup for running with microservices.

Key Features

  • Asynchronous Communication: Enables decoupled, non-blocking inter-service interactions, aside from the unbonding pipeline which follows a different interaction pattern.
  • Fault Tolerance: Utilizes RabbitMQ's message retry mechanism for resilience against transient failures.
  • Horizontal Scalability: Supports increasing system capacity by adding more processing nodes as demand grows.

For more detailed rules applied to message processing, refer to the queue handler documentation.

Workflow

Standard Staking Path

The standard staking path encompasses user-initiated staking through CLI/UI, followed by waiting for the staking period (timelock) to expire.

  1. Transaction Submission: User-submitted staking transactions are confirmed and picked up by the indexer after receiving sufficient Bitcoin block confirmations.
  2. Event Queuing: The indexer sends ActiveStakingEvent messages to the Active Event Queue for the staking API service to process.
  3. Processing and State Management: The staking API service executes statistical calculations, data transformation, and staking state management, inserting records into the timelock_queue collection.
  4. Timelock Expiry Monitoring: A dedicated service monitors the timelock_queue for records with expired Bitcoin Staking timelocks and signals the staking API service to update the staking delegation status to unbonded. This status displays to the user that the staking transaction is ready for withdrawal.

Early Unbonding Path

The early unbonding path enables users to on-demand unlock their staked Bitcoin before the timelock of the staking transaction expires.

  1. Signature Submission: Via the UI/CLI, users can initiate an early unbonding action, which involves generating an unbonding transaction and creating a signature for it, both of which are forwarded to the staking API service.
  2. Signature Verification and Storage: The staking API service validates the signature against the unbonding transaction and stores it for further processing by the unbonding pipeline.
  3. Committee Co-Signing: The unbonding pipeline collects additional signatures from the covenant committee and submits the unbonding transaction to the Bitcoin network.
  4. Transaction Detection: The unbonding transaction is detected by the staking-indexer once it receives a sufficient number of confirmations and a corresponding unbonding event is placed into the RabbitMQ queue.
  5. Processing and State Management: Similar to the standard path, the staking API service handles statistical updates, adjusts the staking state, and inserts a record into the timelock_queue.
  6. Finalization: The expire-service processes items from the timelock_queue in MongoDB and emits an expired event to RabbitMQ for the staking-api-service to process. This marks completed staking transactions as unbonded. This status displays to the user that the staking transaction is ready for withdrawal.

Getting Started

Prerequisites

  • Docker
  • Go

Installation

  1. Clone the repository:
git clone [email protected]:babylonlabs-io/staking-api-service.git
  1. Run the service:
make run-local

OR, you can run as a docker container

make start-staking-api-service
  1. Open your browser and navigate to http://localhost to see the api server running.

Tests

The service only contains integration tests so far, run below:

make tests

Update Mocks

  1. Make sure the interfaces such as the DBClientis up to date
  2. Install mockery: https://vektra.github.io/mockery/latest/
  3. Run make generate-mock-interface

Contribution

Feel free to submit a pull request or open an issue.