Skip to content

wavesplatform/data-service

Repository files navigation

Waves data service API

⚠️ This service is currently in /v0. Breaking changes are coming in /v1 (also possible, but not likely, within /v0 releases). Please use with caution.

This is an API aimed at retrieving data from blockchain quickly and conveniently. We support public APIs for:

Visit /docs for Swagger documentation.

Data service on-premise

It is possible to create your own instance of this service. To do so, follow the guide below.

Requirements

  1. PostgreSQL 11 database with a table stricture found in wavesplatform/blockchain-postgres-sync
  2. Downloaded and continuously updated blockchain data in the database
  3. NodeJS or Docker for either running the service directly, or in a container

Installation and start

The service uses following environment variables:

Env variable Default Required Description
PORT 3000 NO HTTP service port
PGHOST YES Postgres host address
PGPORT 5432 NO Postgres port
PGDATABASE YES Postgres database name
PGUSER YES Postgres user name
PGPASSWORD YES Postgres password
PGPOOLSIZE 20 NO Postgres pool size
PGSTATEMENTTIMEOUT false NO Postgres statement_timeout number in ms. 0 disables timeout, false — use server settings; at this moment used only as default STATEMENT_TIMEOUT
LOG_LEVEL info NO Log level ['info','warn','error']
DEFAULT_MATCHER YES Default matcher public address
MATCHER_SETTINGS_URL NO Default matcher URL for getting settings
DEFAULT_TIMEOUT 30000 NO Default timeout in ms; at this moment used only as PG STATEMENT_TIMEOUT

PGPOOLSIZE is used by the pg-pool library to determine Postgres connection pool size per NodeJS process instance. A good value depends on your server and db configuration and can be found empirically. You can leave it at the default value to start with.

Set those variables to a variables.env file in the root of the project for convenience. In the next steps we will assume this file exists.

If you would like to use some other way of setting environment variables, just replace relevant commands below with custom alternatives.

Docker

If you wish to build data-service image locally, run this command from the project root

docker build -t wavesplatform/data-service .

Otherwise you can use our public image from https://hub.docker.com/r/wavesplatform/data-service

Run the container using this command:

docker run -p=<port>:3000 --env-file=variables.env wavesplatform/data-service

A server will start at localhost:<port> (used in the docker run command). Logs will be handled by Docker. Use any other Docker options if necessary.

When using the container in production, we recommend establishing a Docker logging and restart policy.

NodeJS
  1. Install dependencies
    npm install    # or `yarn install`, if you prefer
  2. Build the server
    npm run build
  3. Start the server
    export $(cat variables.env | xargs) && NODE_ENV=production node dist/index.js

Server will start at localhost:PORT (defaults to 3000). Logs will be directed to stdout.

If you decide to use NodeJS directly (without Docker), we recommend using a process manager, such as pm2.

Daemons

To add candles and pairs functionality the following Docker daemons must be used:

  • Candles — calculate candles for exchange transactions (see description)
  • Pairs — calculate last pairs for 24h exchange transactions (see description)

Documentation

You can run your own instance of Swagger online documentation. To do this, you have to:

  1. Build Docker image from docs/ directory:
    docker build -t wavesplatform/data-service-docs docs/
  2. Run the container
    docker run --rm -d -p 8080:8080 -e SWAGGER_JSON=/app/openapi.json wavesplatform/data-service-docs

Its will start the documentation server at localhost:8080. Enjoy!

General recommendations

  • Set up a dedicated web server such as Nginx in front of data-service backends (for ssl/caching/balancing);
  • Implement a caching strategy. Different endpoints may need different cache time (or no cache at all);
  • Run several process instances behind a load balancer per machine. docker-compose --scale can help with that, or it can be done manually. A good rule of thumb is to use as many instances as CPU cores available;
  • Use several machines in different data centers and a balancer to minimize downtime;
  • Experiment with PostgreSQL settings to find out what works best for your configuration. Tweaking PGPOOLSIZE also can help performance;
  • Run the sql from mainnet.sql to increase exchange-transactions service performance.