Turn this feature off for now & re-implement once branch set to public

The builder stage packages all the tools and compile-time dependencies for
your application. A mix release will be copied in the
running stage.

Automate the pushing of your Docker image to the GitHub Docker registry if
your CI pipeline passes. Authentication with the registry is required before
being able to push. You should create three tags for your Docker image: one
matching the commit it refers to, one matching the PR number (if present),
and other that is "latest".

Create a Docker Compose file that includes a PostgreSQL database and your
Phoenix Live View application.

Add a new GitHub action to your CI/CD pipeline that automatically deploys
your Phoenix Live View application any time any code is merged into the main
branch of your repository. This action should use the build tags that are
created in the CI pipeline.

Currently, the dayshours string that s returned from SF food truck data has multiple patterns. Write a function to extract significant bits of data to improve search.

Make the different instances of your Phoenix Live View Application form part
of a distributed Erlang cluster. Do this using the libcluster library after
you start the application as an Erlang node.

Use Packer to create a custom AMI for your EC2 instances and then link it
to your existing AWS Terraform infrastructure. This AMI should pre-install
Docker and netcat.

Adapt your AWS Terraform configuration to use an Auto Scaling Group so that
your Elixir application scales up/down depending on the average CPU usage of
the EC2 intances in that group.

Create an alert so that you are notified any time the average CPU of your
Elixir applications reaches above 50%.

Use Docker secrets to hide the values to the following variables in your
Docker Compose file: DATABASE_URL, POSTGRES_PASSWORD, and SECRET_KEY_BASE.

Revise your Terraform configuration so that it can support a multi-node
swarm. To do this, create more than one EC2 instance and distribute it
across 3 availibility zones.

The PostGIS adds storing, indexing, and querying geo-spatial data support to Postgres.
Currently, there is a blocker on ARM chips which is preventing an asdf Postgres installation
from creating a PostGIS extension.

Links and References

PostGIS M1 Issue

the docker postgis image doesn't seem to support arm chips

asdf-postgres issue #5 discussing how to install postgis for an asdf postgres install

1. Discuss -> open source project. Create a repo on Github, use git for source control, and use a Readme file to document what you built for the newcomer to your project.
2. Document how to get started in terms of setup so EL would be able to run it.

Update the deployment keys of your Phoenix Live View application so that
it automatically rolls back to a previous version if the image you want
to deploy does not start properly.

Target: add a point db type to food_trucks table/schema
→ point: %Geo.Point{coordinates: {truck.lng, truck.lat}, srid: 4326}

Currently, the fooditems string that is returned the SF food truck data has a few differing patterns. Map this string to improve search.

Add all of the AWS resources that will allow you to deploy your application
to a remote single-node to your Terraform configuration. (Making the
resources support a multi-node will be in another task). Required: an EC2
instance, a VPC, a security group, and an SSH key pair.

Once the dynamic_private_key.pem key has been ssh-added and the currently running instance's IP has been written to your local known_hosts file, use the DOCKER_HOST env var to connect to an iex session.

example:

$ ssh-add ./environments/production/dynamic_private_key.pem

$ ssh-keyscan -H 54.193.68.139 >> ~/.ssh/known_hosts

connect to an iex> REPL

$ DOCKER_HOST=ssh://[email protected] iex -S mix

and seed some data

iex> BellyBot.GlobalSetup.seed_food_trucks()

# check the food truck count
iex> BellyBot.FoodTrucks.list_food_trucks() |> Enum.count()

ISSUE:
The data gets seeded but is not showing up in the application. Perhaps the docker instance and volume are not communicating? More investigation needed.

Implement a CI pipeline that includes all of the necessary steps when
delivering an Elixir application: code compilation, dependency caching,
testing, code formatting, an unused dependency check.

Once an EC2 instance is deployed and available to navigate to, ec2-54-193-68-139.us-west-1.compute.amazonaws.com:4000 -> 2500k network panel errors ensue.

GET http://ec2-54-193-68-139.us-west-1.compute.amazonaws.com:4000/live/longpoll?_csrf_token=ExEsFRwaeBU7FBAJIDsoajNCaT1vKQtRrKaDxTUpcmcALcC8Gz0_XCs8&_track_static[0]=http://ec2-54-193-68-139.us-west-1.compute.amazonaws.com:4000/assets/app-a86ae2aa16a2e20e835a380067392207.css?vsn=d&_track_static[1]=http://ec2-54-193-68-139.us-west-1.compute.amazonaws.com:4000/assets/app-a7b7a1bb2570182b93179aee6863d1bb.js?vsn=d&_mounts=0&_live_referer=undefined&vsn=2.0.0&=

Currently, running seeds populate the food truck data, once. This will not suffice. Add Oban and build a nightly task to fetch and refresh the SF food truck data.

Use Loki, Promtail agent to collect the logs for your Docker containers and
Grafana to display them. Then use Prometheus/PromEx to collect the
metrics exposed by the main libraries used in an Elixir application.

This stage copies the release built in the builder stage and uses it as the
entrypoint of your Docker image with the minimum system requirement to run
it.

Build out an MVP Phoenix Live View App.

• Implement seeds to fetch food truck data
• Implement list of food trucks
• Implement a show page for each truck
  Implement basic search with flop