@uma/bot-strategy-runner

@uma/bot-strategy-runner

This package contains a generalized bot runner enabling simple execution of multiple concurrent UMA bots on one machine.

Installing the package

yarn add @uma/bot-strategy-runner
yarn build

Or, if you cloned from the mono repo simply run the following from the root of the mono repo:

yarn
yarn qbuild

Quick start

This package requires a simple JSON config to parameterize the strategy runner. A simple config that enables a liquidator and disputer on 3 contracts would look like this:

{
  "globalAddressWhitelist": [
    "0x0f4e2a456aAfc0068a0718E3107B88d2e8f2bfEF",
    "0xd9af2d7E4cF86aAfBCf688a47Bd6b95Da9F7c838",
    "0xd60139B287De1408f8388f5f57fC114Fb4B03328"
  ],
  "commonConfig": {
    "priceFeedConfig": {
      "cryptowatchApiKey": "YOUR-CRYPTO-WATCH-API-KEY"
    }
  },
  "liquidatorSettings": { "enableBotType": true },
  "disputerSettings": { "enableBotType": true },
  "monitorSettings": { "enableBotType": false }
}

Add this contents to a BotRunnerConfig.json file within this package.

Then, to start the strategy runner run:

node ./dist/src/index.js --fileConfig ./BotRunnerConfig.json

This will spin up liquidators and monitors on these three contract addresses.

All config options

The strategy runner is design to support a wide range of execution configurations with the goal of being as configurable as standalone UMA bots, while enabling paralization. Some of these configs are discussed below.

Reading in configs from remote github

Bot runners might want to store their config files within a github repo. This can be read in as follows:

node ./dist/src/index.js --urlConfig  https://api.github.com/repos/<your-organization>/<your-repo-name>/contents/<path-to-your-json-file> -accessToken <github-access-token>

Be sure to replace the <your-x> content with your actual values. This kind of config can be used in conjunction with the --fileConfig flag, wherein the configs will be joined.

Config file structure and custimization

The config file can be customized with a bunch of extra params and settings. The snippet below showcases each section and what can be placed in it for a fully custom setup.

interface strategyRunnerConfig {
  botNetwork?: string; // network to connect the executed bots on. Defaults to `mainnet_mnemonic`
  pollingDelay?: number; // how frequently the strategy runner should re-run after all strategies have finished. defaults to 120 seconds. If set to 0 the strategy runner will run in serverless mode.
  botConcurrency?: number; // how many strategies should be run in parallel. defaults to 10.
  strategyTimeout?: number; // how long the runner should wait if a strategy is timing out. Defaults to 60 seconds.
  verboseLogs?: boolean; // toggles if the runner should produce all logs generated by all strategies. WARNING: can be noisy! Defaults to false.
  emitDebugLogs?: boolean; // toggles if the runner should produce debug logs in execution output. If set to false, the runner will produce any info and above logs on execution. Defaults to false.
  globalAddressWhitelistUrls?: Array<string>; // define an array of URLS that the runner should pull configs in from. UMA maintains a whitelist of approved dev mining contracts that acts as a good starting point for a whitelist of UMA contracts to run bots on.
  globalAddressWhitelist?: Array<string>; // define array of addresses to run strategies on.
  globalAddressBlacklist?: Array<string>; // define array of addresses to blacklist the strategies strategies on. Used to exclude some addresses from URL whitelists.
  commonConfig?: { [key: string]: any }; // define any common configs to apply to all bot types.
  liquidatorSettings?: botSettings; // instance of botSettings. See code snippet below for params.
  disputerSettings?: botSettings;
  monitorSettings?: botSettings;
}

Each bot type can be configured with specific settings for that bot as well as additional overrides and settings. The example below is of type botSettings and would be placed within the previous config under the botSettings section for each bot type.

interface botSettings {
  enableBotType: boolean; // boolean to enable or disable a bot type. By default is set to false (disabled)
  addressWhitelist?: Array<string>; // add specific whitelists for this particular bot type. For example if you want a specific address to only have the liquidator running then this would be used.
  addressBlacklist?: Array<string>; // add specific blacklists for this particular bot type. For example if you dont want to run a disputer on a particular address you would add this here.
  commonConfig?: { [key: string]: any }; // Add custom config for all bots of this type. For example if you want all liquidators to have a specific config you should add it here.
  addressConfigOverride?: { [key: string]: { [key: string]: any } }; // override a particular config for a specific contract address. This enables address specific granularity in the configs used for different bot types.
}