Migrating to Shuttle

​

Setup

This guide assumes that you have Shuttle already setup. If you don’t, head on over to our installation & quick start guides in order to install Shuttle & create your first project.

For this example, we’ll assume you are migrating your Axum project. Once you’ve created your project, you’ll want to add shuttle_runtime and shuttle_axum to your dependencies to be able to use shuttle’s runtime with the Axum framework. You can run the following command below like so to do this:

cargo add shuttle_runtime shuttle_axum

Any secrets you need to use will be kept in a Secrets.toml file (dev secrets to be held in Secrets.dev.toml) which will be placed at the Cargo.toml level - you’ll want to use the shuttle_secrets crate to be able to use the new secrets:

cargo add shuttle_secrets

Your secrets file will look the same as your previous .env file, so you won’t need to change anything.

You can also easily get a provisioned database like so (this example will be for a provisioned PostgresQL instance specifically):

cargo add shuttle-shared-db --features postgres

If you have any database records you’d like to keep, it would be a good idea to make sure you export them before you start as you’ll want to make sure those are preserved; otherwise, you’ll lose the data. You will not need a secrets file if you only need a provisioned PG database - this will be automatically be provisioned to you in the form of a SQLx connection.

​

Migrating your Code

To be able to run your project on Shuttle, you need to make a few changes to your code. Instead of running on the tokio runtime, you should use the shuttle_service::main macro instead like so and swap out dotenvy for Shuttle’s Postgres annotation:

This is what your main.rs file looks before:

// main.rs
#[tokio::main]
async fn main() {
    dotenv().ok();

    let postgres = dotenvy::var("DATABASE_URL").expect("No database URL was set!");

    let postgres = sqlx::Pool::connect(&postgres).await.unwrap();

    sqlx::migrate!()
        .run(&postgres)
        .await
        .expect("Migrations failed :(");

    let router = create_api_router(postgres);
    let addr = SocketAddr::from(([0, 0, 0, 0], 8000));

    Server::bind(&addr)
        .serve(router.into_make_service())
        .await
        .unwrap()
}

And this is what it looks like after:

#[shuttle_runtime::main]
pub async fn axum (
#[shuttle_shared_db::Postgres] postgres: PgPool,
#[shuttle_secrets::Secrets] secrets: shuttle_secrets::SecretStore
) -> shuttle_axum::ShuttleAxum {

    sqlx::migrate!()
        .run(&postgres)
        .await
        .expect("Migrations failed :(");

 let router = create_api_router(postgres);

Ok(router.into())
}

If you need more than a simple router, you’ll want to create a custom struct that holds all of your required app state information inside and then create an impl for the struct - you can find more about that here. Anything outside of your entry point function (the function that uses the shuttle_runtime::main macro) doesn’t need to be changed. If you are using secrets as well as a database connection, you may wish to create a struct that holds both of these values and then pass it into the function that generates the router.

​

Deploying

If you haven’t yet, you will probably want a Shuttle.toml file at the Cargo.toml level to name your project to whatever you’d like, like so:

name="<app-name-here>"

Now all you need to do is to run the following commands:

cargo shuttle project new
cargo shuttle project deploy

Your project should now be deployed!