bench-bot

This is a substrate-only bot at the moment.

How to use

The bot runs commands in response to pull request comments (example). The form is:

/bench [action] [...args]

Kilt bench Command	Chain Spec	Weight template
/bench runtime pallet {pallet_name}	dev	pallets/{pallet_folder}/src/default_weights.rs
/bench runtime spiritnet-pallet {pallet_name}	spiritnet-dev	pallets/{pallet_folder}/src/default_weights.rs
/bench runtime peregrine {pallet_name}	dev	runtimes/peregrine/src/weights/{pallet_name}.rs
/bench runtime spiritnet-runtime {pallet_name}	spiritnet-dev	runtimes/spiritnet/src/weights/{pallet_name}.rs

Naming convention: If pallet folder contains hyphen, output file inside runtime will be folder name replaced by underscore otherwise the same. You can manually run scripts inside mashnet-node/scripts directory to run all runtime pallets.

Environment variables and Github settings have to properly configured upfront for this interaction to work.

How to restart

sudo /home/benchbot/bench-bot/run restart

Configuration

Create a .env file in the root with the following:

APP_ID=<App id from Github App Settings>
BASE_BRANCH=<the default branch for merging into the PRs, e.g. master>
CLIENT_ID=<Client ID from Github App Settings>
CLIENT_SECRET=<Client ID from Github App Settings>
PRIVATE_KEY_PATH=<Path the the private key of the github app>
WEBHOOK_SECRET=<Webhook secret from Github App Settings>
WEBHOOK_PROXY_URL=<optional; webhook proxy for development>

During development it's recommended to use smee for WEBHOOK_PROXY_URL because it enables testing your bot's functionality locally, without having to SSH into the dedicated machine.

Linting and formatting

The commands yarn run format and yarn run lint are available for ensuring style consistency on this project's code.

Running

Locally

yarn && yarn start

Dedicated machine

Note: Before disrupting the production deployment, it's first recommended to check if some benchmark is running through pgrep -a cargo .

The run script is used to manage the application. Use run help for documentation about its options.

run bootstrap will take care of creating and installing everything from scratch. After install, you'll also need to set up environment variables which optionally can be done through a .env file in the bot's installation path.

Bot commands

• run {start,stop,restart}: execute the relevant action for the bot.
• run update [ref]: restart the bot with the branch or PR
  • For branch: ssh user@remote '/home/benchbot/bench-bot/run update master'
  • For PR: ssh user@remote '/home/benchbot/bench-bot/run update pull/number/head:branch' e.g. pull/1/head:master

Logs

See https://gitlab.parity.io/groups/parity/opstooling/-/wikis

Required Github settings

Permissions

• Metadata: Read Only
• Issues: Read/Write
• Pull Requests: Read/Write
• Contents: Read/Write

Event subscriptions

• Issue comments