UMA bot automator

This project automates the deploy and provisioning of a UMA liquidation bot on Google Cloud.

Tools

Terraform provide the structure at the GCP.
Ansible execute playbooks and install required softwares over the VM.
Docker abstract all tools and don't blow up your localhost.

Prerequisites

Docker

Getting started

Setup

First off all, clone this repository and setup the docker image with all the tools to get things done.

$ git clone https://github.com/evaldofelipe/uma-automator
$ cd uma-automator
$ make setup

This process take a wile. Use this time to create the GCP project at console.

New at Google Cloud? Follow these steps!

If you don't have a GCP account yet, you could use the the free trial provide by google to execute this project. You can create your free account here.
After create the account ready, you need create a project. Follow this documentation to guide you.

IMPORTANT: Choose a project name as you want, and store the project_ID not the project name! You'll use later on the project.

After the project was created, we need interact with GCP using a Service Account, an key file provide to use cloud API's. Follow this guide to create the Service Account, with Owner/Owner permissions, download the json file (the create key button are just a bit hide!), and store the file on a security place. Ansible and terraform will use this key to interact with GCP.
On this automation are possible to choose the region where your VM are deployed, but as a new GCP user, I recommend, don't worry about it, and use the default region already set here.

Already GCP user? Come on this way!

At this time, this project don't support the project automatic creation, because the trial account don't have a main organization to control the projects. So it's necessary create a new project on the console.
Create the project and store and store the project_ID. You'll use later on the project.
Create a Service Account on the new project, with Owner/Owner permissions, download the json, and store the file on a security place. Ansible and terraform will use this key to interact with GCP.
If you want change the default Region and Zones defined here, just keep in mind to don't forget to select the right variables. If you need, take a look on the regions list.

Finish the local setup

IMPORTANT: On this automation we assume your ssh keys are stored at $HOME/.ssh/ with the names id_rsa and id_rsa.pub if you're using other path, edit the path at Makefile on line 19:

--volume $(HOME)/.ssh/:/home/uma/.ssh/:ro \

With a new project, and their respective Service Account, export your SA as a env variable, and the project ID.

$  export GCP_TOKEN=$(cat /path/to/your/key.json)
$  export PROJECT="your-project-ID"

You will need export the variables used by the bot descrided this documentation on the file example.env. The automation will make sure that file will be created on the right place.

$ export EMP_ADDRESS="your-emp-address"
$ export PRIVATE_KEY="your-private-key"
$ export CUSTOM_NODE_URL="your-custom-node-url"
$ export COMMAND="yarn --silent liquidator --network mainnet_privatekey"

After this modifications, you're good to go!

Deploy resources

creating the infrastructure

To create the environment, execute:

$ make build-all

Provisioning the bot

After the VM was deployed, configure executing:

$ make provision-all

Check service status

You could check if your bot was deployed on the right way cheking the current logs executing:

$ make logs-check

After that, you can download the log file for your localhost:

$ make logs-download

Destroy the infrastructure

After realize your tests, you can destroy everything just running:

$ make destroy-all