Grouparoo `app-example-gcp`

An example Grouparoo project for deploying Grouparoo on Google Cloud Platform (GCP) with Node.JS

Goal: To create Grouparoo deployment that:

Relies on Google's hosted databases for persistence.
Is process-managed locally

Limitations:

This quick configuration does not have application high-availability, as it relies on a single host. This is not a 12-factor app with configuration stored in the Environment
Will not re-deploy as your code changes. You can do that manually, or use a tool like Chef or Ansible.

How we got here

Create a new Grouparoo project. Learn more @ https://www.grouparoo.com/docs/installation.

npm install -g grouparoo
grouparoo init .

Install the Grouparoo plugins you want, e.g.: grouparoo install @grouparoo/postgres. Learn more @ https://www.grouparoo.com/docs/installation/plugins

Running this Repo

Assuming you have node.js installed (v12+):

git clone https://github.com/grouparoo/app-example-gcp.git
cd app-example-gcp
npm install
cp .env.example .env
npm start

Deployment Steps

Configure a new GCP project for Grouparoo, or choose an existing one
Create the Grouparoo Database
- Choose Postgres CLoud SQL
- A standard server should be OK, we recommend 4 vCPU, 15 GB
- Enable the Private IP address & an Automatically Assigned IP Range.
- Once the database is running, create a new database like grouparoo_production
- Note and build DATABASE_URL in a scratch document. The default username is postgres.
Create the Grouparoo Redis Server
- Choose Memorystore - Redis
- Create 1 replica
- 1GB or capacity should be enough
- "Enable Auth" but disable "in transit encryption" for redis
- Use the Private IP address & an Automatically Assigned IP Range settings you created for Postgres above.
- Once redis is up and running, note the AUTH string. You'll use this as the "password" in REDIS_URL. The username should be a 1-char letter which signals it should be ignored, like "r"
- Note and build REDIS_URL in a scratch document
Create a Compute Engine Server to run your Grouparoo Instances
- Choose a machine with at least 2CPUs and 4GB of memory
- Choose the Ubuntu 20.04 LTS Minimal Boot Disk
- Enable the HTTP and HTTPS ports in the firewall
- ssh into the instance from the console, and:

# Update packages
sudo apt-get update && sudo apt-get upgrade -y

# Install Tools
sudo apt-get install iptables net-tools vim git -y

# Install NVM and Node.JS
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.38.0/install.sh | bash
source ~/.bashrc
nvm install v14

# Git clone/copy your code to the server
# For example...
git clone https://github.com/grouparoo/app-example-gcp
cd app-example-gcp

# << be sure that you are in your application directory >>

# Install packages
npm install

# In your project, copy and configure the .env file to match your environment
cp .env.example .env
vim .env # Fill in your server information; enable both WEB and WORKERS

# Configure an internal forward from port 3000 (the app) to port 80
# Note: your ethernet interface may not be `ens4`, it may be something else.  Run `ifconfig -a` to see your network interfaces
sudo iptables -A INPUT -i ens4 -p tcp --dport 80 -j ACCEPT
sudo iptables -A INPUT -i ens4 -p tcp --dport 3000 -j ACCEPT
sudo iptables -A PREROUTING -t nat -i ens4 -p tcp --dport 80 -j REDIRECT --to-port 3000

# install iptables-persistent to save your rules
sudo apt-get install iptables-persistent -y # answer 'yes' to save your existing iptables rules

# prepare logging directory for access
sudo mkdir /var/log/grouparoo
sudo chown -R $USER /var/log/grouparoo

Notes on configuring your .env on the server:

Keep the PORT=3000. We've configured iptables to forward all traffic on port 80 to 3000 internally. This means our unprivileged app user can doesn't need to run the app as sudo.
- The WEB_URL will be something like WEB_URL=http://1.2.3.4
For this example, we'll be making a single process which acts as both web and background workers (WORKERS=10 and WEB_SERVER=true)
Use the values from the new Google Cloud services for DATABASE_URL and REDIS_URL

Now, we'll use pm2 to monitor and run the Grouparoo application:

# << be sure that you are in your application directory >>

# Install PM2
npm install -g pm2

# Install PM2 into your OS's init system
pm2 startup # learn more @ https://pm2.keymetrics.io/docs/usage/startup/
# Follow the prompts to register pm2 with systemd

# start the grouparoo process
pm2 start --name grouparoo --interpreter=bash node_modules/@grouparoo/core/bin/start

# see process information
pm2 list # see running processes
pm2 logs # follow process logs (ie --tail)
pm2 monit # fancy process monitor

Configure Monitoring
Configure Load Balancer, SSL, and DNS

Updating the Deployment

SSH to your server

# in the application directory
git pull
npm install
pm2 restart grouparoo

Notes

In this example, we deployed a single server type that acts as both a WEB and BACKGROUND WORKER service for simplicity. For more robust deployments, you may want to split these tasks to different servers.

You may want to modify logging behavior with:

GROUPAROO_LOGS_STDOUT_DISABLE_TIMESTAMP=true- Don't prepend timestamps to log messages
GROUPAROO_LOGS_STDOUT_DISABLE_COLOR=true- Disable colorizing log message lines
GROUPAROO_LOGS_PATH="/var/log/grouparoo"- Logs are stored in /var/log/grouparoo with this setting enabled. You can store them wherever you like.

Visit https://github.com/grouparoo/app-examples to see other Grouparoo Example Projects.

License

grouparoo/app-example-gcp

ErrorLooks like something went wrong!

About

Resources

License

Security policy

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Contributors 3

Uh oh!

Languages

Error
Looks like something went wrong!

Packages