wasp/web/docs/deploying.md

title
Deploying

import useBaseUrl from '@docusaurus/useBaseUrl';

import AddExternalAuthEnvVarsReminder from './_addExternalAuthEnvVarsReminder.md'

:::info Wasp is in beta, so keep in mind there might be some kinks / bugs, and possibly a bit bigger changes in the future. If you encounter any issues, reach out to us on Discord and we will make sure to help you out! :::

Automated

Wasp CLI

Using the Wasp CLI, you can easily deploy a new app to Fly.io with just a single command:

wasp deploy fly launch my-wasp-app mia

Under the covers, this runs the equivalent of the following commands:

wasp deploy fly setup my-wasp-app mia
wasp deploy fly create-db mia
wasp deploy fly deploy

In the above commands, we used an app basename of my-wasp-app and deployed it to the Miami, Florida (US) region (called mia). The basename is used to create all three app tiers, so you will have three components in your Fly dashboard:

• my-wasp-app-client
• my-wasp-app-server
• my-wasp-app-db

:::caution Your app name must be unique across all of Fly or deployment will fail. Additionally, please do not CTRL-C or exit your terminal as these commands run. :::

:::note Fly has free allowances for up to 3 VMs. If you already have some apps deployed on their free Hobby Plan, you will need to add your credit card info before proceeding. :::

The list of available Fly regions can be found here. You can also run wasp deploy fly cmd platform regions --context server.

Commands

setup will create your client and server apps on Fly, and add some secrets, but does not deploy them. We need a database first, which we create with create-db, and it is automatically linked to your server.

:::note We only run the setup and create-db steps once.

You may notice after running setup you have a fly-server.toml and fly-client.toml in your Wasp project directory. Those are meant to be version controlled. If you want to maintain multiple apps, you can add the --fly-toml-dir <abs-path> option to point to different directories, like "dev" or "staging". :::

Finally, we deploy which will push your client and server live. We run this single command each time you want to update your app.

:::note Fly.io offers support for both locally built Docker containers and remotely built ones. However, for simplicity and reproducability, the CLI defaults to the use of a remote Fly.io builder. If you wish to build locally, you may supply the --build-locally option to wasp deploy fly launch or wasp deploy fly deploy. :::

If you would like to run arbitrary Fly commands (eg, flyctl secrets list for your server app), you can run them like so:

wasp deploy fly cmd secrets list --context server

:::note If you are deploying an app that requires any other environment variables (like social auth secrets), you will want to set your environment variables up like so:

wasp deploy fly cmd secrets set GOOGLE_CLIENT_ID=<...> GOOGLE_CLIENT_SECRET=<...> --context=server

:::

Manual

In addition to the CLI, you can deploy a Wasp project by generating the code and then deploying generated code "manually", as explained below.

In the future, the plan is to have Wasp take care of it completely: you would declaratively define your deployment in .wasp and then just call wasp deploy (github issue).

If you want to deploy your App completely free of charge, continue reading below for guides on using Fly.io as your backend (server) provider and Netlify for your frontend (client).

If you prefer to host client and server on one platform, and don't mind paying a very small fee for extra features, we suggest following the guide on using Railway as your provider.

Generating deployable code

wasp build

generates deployable code for the whole app in the .wasp/build/ directory. Next, we will deploy this code.

NOTE: You will not be able to build the app if you are using SQLite as a database (which is a default database) -> you will have to switch to PostgreSQL.

Deploying API server (backend)

In .wasp/build/, there is a Dockerfile describing an image for building the server.

To run server in production, deploy this docker image to your favorite hosting provider, ensure that env vars are correctly set, and that is it.

Below we will explain the required env vars and also provide detailed instructions for deploying to Fly.io or Heroku.

Env vars

Server uses following environment variables, so you need to ensure they are set on your hosting provider:

• PORT -> The port number at which it will listen for requests (e.g. 3001).
• DATABASE_URL -> The URL of the Postgres database it should use (e.g. postgresql://mydbuser:mypass@localhost:5432/nameofmydb).
• WASP_WEB_CLIENT_URL -> The URL of where the frontend app is running (e.g. https://<app-name>.netlify.app), which is necessary for CORS.
• JWT_SECRET -> You need this if you are using Wasp's auth feature. Set it to a random string (password), at least 32 characters long.

Deploying to Fly.io (free, recommended)

Fly.io offers a variety of free services that are perfect for deploying your first Wasp app! You will need a Fly.io account and the flyctl CLI.

:::note Fly.io offers support for both locally built Docker containers and remotely built ones. However, for simplicity and reproducability, we will default to the use of a remote Fly.io builder.

Additionally, fly is a symlink for flyctl on most systems and they can be used interchangeably. :::

Make sure you are logged in with flyctl CLI. You can check if you are logged in with flyctl auth whoami, and if you are not, you can log in with flyctl auth login.

Set up a Fly.io app (only once per Wasp app)

Unless you already have a Fly.io app that you want to deploy to, let's create a new Fly.io app. Position yourself in .wasp/build/ directory (reminder: which you created by running wasp build previously):

cd .wasp/build

Now from within the build directory, run the launch command to set up a new app and create a fly.toml file:

flyctl launch --remote-only

This will ask a series of questions, including what region to deploy in and if you would like a database.

• Say yes to "Would you like to set up a Postgresql database now?", and select Development, and Fly.io will set a DATABASE_URL for you.
• Say no to "Would you like to deploy now?", as well as any additional questions. We still need to set a few environment variables.

:::note If your attempts to initiate a new app fail for whatever reason, then you can run flyctl apps destroy <app-name> before trying again.

What does it look like when your DB is deployed correctly?

When your DB is deployed correctly, you will be able to view it in the Fly.io dashboard:

:::

Next, let's copy the fly.toml file up to our Wasp project dir for safekeeping.

cp fly.toml ../../

Next, let's add a few more environment variables:

flyctl secrets set PORT=8080
flyctl secrets set JWT_SECRET=<random_string_at_least_32_characters_long>
flyctl secrets set WASP_WEB_CLIENT_URL=<url_of_where_frontend_will_be_deployed>

# If you are using an external auth method (Google or GitHub), make sure to add their vars too!
# flyctl secrets set GOOGLE_CLIENT_ID=<google_client_id>
# flyctl secrets set GOOGLE_CLIENT_SECRET=<google_client_secret>

NOTE: If you do not know what your frontend URL is yet, don't worry. You can set WASP_WEB_CLIENT_URL after you deploy your frontend.

If you want to make sure you've added your secrets correctly, run flyctl secrets list in the terminal. Note that you will see hashed versions of your secrets to protect your sensitive data.

Deploy to a Fly.io app

While still in the .wasp/build/ directory, run:

flyctl deploy --remote-only --config ../../fly.toml

This will build and deploy the backend of your Wasp app on Fly.io to https://<app-name>.fly.dev! 🤘🎸

Now, if you haven't, you can deploy your frontend -- we suggest using Netlify for this -- and add the client url by running flyctl secrets set WASP_WEB_CLIENT_URL=<url_of_deployed_frontend>

Additionally, some useful commands include:

flyctl logs
flyctl secrets list
flyctl ssh console

Redeploying after Wasp builds

When you rebuild your Wasp app (with wasp build), it will remove your .wasp/build/ directory. In there, you may have a fly.toml from any prior Fly.io deployments. While we will improve this process in the future, in the meantime, you have a few options:

1. Copy the fly.toml file to a versioned directory, like your Wasp project dir. From there, you can reference it in flyctl deploy --config <path> commands, like above.
2. Backup the fly.toml file somewhere before running wasp build, and copy it into .wasp/build/ after. When the fly.toml file exists in .wasp/build/ dir, you do not need to specify the --config <path>.
3. Run flyctl config save -a <app-name> to regenerate the fly.toml file from the remote state stored in Fly.io.

Deploying to Heroku (non-free)

:::note Heroku used to offer free apps under certain limits. However, as of November 28, 2022, they ended support for their free tier. https://blog.heroku.com/next-chapter

As such, we recommend using an alternative provider like Fly.io for your first apps. :::

You will need Heroku account, heroku CLI and docker CLI installed to follow these instructions.

Make sure you are logged in with heroku CLI. You can check if you are logged in with heroku whoami, and if you are not, you can log in with heroku login.

Set up a Heroku app (only once per Wasp app)

Unless you already have a heroku app that you want to deploy to, let's create a new Heroku app:

heroku create <app-name>

Unless you have external Postgres database that you want to use, let's create new database on Heroku and attach it to our app:

heroku addons:create --app <app-name> heroku-postgresql:mini

:::caution

Heroku does not offer a free plan anymore and mini is their cheapest database instance - it costs $5/mo.

:::

Heroku will also set DATABASE_URL env var for us at this point. If you are using external database, you will have to set it yourself.

The PORT env var will also be provided by Heroku, so the only two left to set are the JWT_SECRET and WASP_WEB_CLIENT_URL env vars:

heroku config:set --app <app-name> JWT_SECRET=<random_string_at_least_32_characters_long>
heroku config:set --app <app-name> WASP_WEB_CLIENT_URL=<url_of_where_frontend_will_be_deployed>

NOTE: If you do not know what your frontend URL is yet, don't worry. You can set WASP_WEB_CLIENT_URL after you deploy your frontend.

Deploy to a Heroku app

Position yourself in .wasp/build/ directory (reminder: which you created by running wasp build previously):

cd .wasp/build

assuming you were at the root of your Wasp project at that moment.

Log in to Heroku Container Registry:

heroku container:login

Build the docker image and push it to Heroku:

heroku container:push --app <app-name> web

App is still not deployed at this point. This step might take some time, especially the very first time, since there are no cached docker layers.

:::note

Note for Apple M1 users

Apple M1 users need to build a non-Arm image, so the above step will not work at this time. Instead of heroku container:push, users instead should:

docker buildx build --platform linux/amd64 -t <app-name> .
docker tag <app-name> registry.heroku.com/<app-name>/web
docker push registry.heroku.com/<app-name>/web

You are now ready to proceed to the next step. :::

Deploy the pushed image and restart the app:

heroku container:release --app <app-name> web

This is it, backend is deployed at https://<app-name>.herokuapp.com!

Additionally, you can check out the logs with:

heroku logs --tail --app <app-name>

:::note

Note on using pg-boss with Heroku

If you wish to deploy an app leveraging Jobs that use pg-boss as the executor to Heroku, you need to set an additional environment variable called PG_BOSS_NEW_OPTIONS to {"connectionString":"<REGULAR_HEROKU_DATABASE_URL>","ssl":{"rejectUnauthorized":false}}. This is because pg-boss uses the pg extension, which does not seem to connect to Heroku over SSL by default, which Heroku requires. Additionally, Heroku uses a self-signed cert, so we must handle that as well.

• https://devcenter.heroku.com/articles/connecting-heroku-postgres#connecting-in-node-js :::

Deploying web client (frontend)

Position yourself in .wasp/build/web-app directory (reminder: which you created by running wasp build previously):

cd .wasp/build/web-app

assuming you were at the root of your Wasp project at that moment.

Run

npm install && REACT_APP_API_URL=<url_to_wasp_backend> npm run build

:::info NO SLASH Make sure your API URL does not have a trailing "/" on the end of it:
✅ https://backend.example.com
❌ https://backend.example.com/ :::

where <url_to_wasp_backend> is url of the wasp backend that you previously deployed, e.g. https://wasp-test.fly.dev.

This will create build/ directory, which you can deploy to any static hosting provider. Check instructions below for deploying to Netlify.

Deploying to Netlify

Netlify is a static hosting solution that is free for many use cases. You will need Netlify account and netlify CLI installed to follow these instructions.

Make sure you are logged in with netlify CLI. You can check if you are logged in with netlify status, and if you are not, you can log in with netlify login.

While positioned in .wasp/build/web-app/ directory, and after you have created .wasp/build/web-app/build/ directory as per instructions above, run

netlify deploy

and carefully follow their instructions (i.e. do you want to create a new app or use existing one, team under which your app will reside, ..., final step to run netlify deploy --prod).

That is it!

NOTE: Make sure you set this URL as the WASP_WEB_CLIENT_URL environment variable in your server hosting environment (e.g., Fly.io or Heroku).

Deploying to Railway ("freemium", all-in-one solution)

Railway makes it easy to deploy your entire app -- database, server, and client -- on one platform. You can use the platform for free for a limited time (~21 days) per month. Upgrading to the Developer plan will only cost you a few dollays per month per service.

:::danger 🛑 Due to Railway's current proxy configuration, Google Auth will not currently work. If you're using Google Auth in your Wasp App, you can still deploy your back-end to Railway, but we suggest you deploy your front-end to Netlify :::

To get started, follow these steps:

1. Generate deployable code (wasp build)
2. Sign up at Railway.app (Tip! Sign up with your GitHub account for $5/month of usage free)
3. Before creating a new project, install the Railway CLI by running the following command in your terminal:

curl -fsSL https://railway.app/install.sh | sh

4. While still in the terminal, run railway login and a browser tab will open to authenticate you.

Create New Project

Go back to your Railway dashboard, click on + New Project, and select Provision PostgreSQL from the dropdown menu.

Once it initializes, right click on the + New button in the top right corner and select >_ Empty Service. Once it initializes, click on it, go to Settings > General and change the name (e.g. server).

Go ahead and create another empty service and name it (e.g. client).

Just in case, here is a helpful screenshot ;)

Deploy to services

Now go back to your terminal and execute the following commands:

1. Move into your app's .wasp/build/ directory, which was created when you ran wasp build previously:

cd .wasp/build

2. "Link" your app build to your newly created Railway project:

railway link

3. Push and deploy the project to railway (make sure you're in .wasp/build):

railway up

Select server when prompted with Select Service. Press enter. Railway will now locate the Dockerfile and deploy your server 👍

When deployment is finished, you might see: Deployment live at <url_to_wasp_backend>. If not, go now to your Railway dashboard and in the server instance's Settings tab, click Generate Domain. Copy the new URL as we will need it for step 5! 📜

4. Next, change into your app's frontend build directory .wasp/build/web-app:

cd web-app

5. Create the production build, adding the URL from step 3:

npm install && REACT_APP_API_URL=<url_to_wasp_backend> npm run build

:::info NO SLASH Make sure your API URL does not have a trailing "/" on the end of it:
✅ https://backend.example.com
❌ https://backend.example.com/ :::

6. Change into the .wasp/build/web-app/build directory:

cd build 

7. Next, we want to link this specific frontend directory to our project as well:

railway link 

8. Deploy the client and select client when prompted with Select Service:

railway up

9. Your apps are deployed 🧙‍♂️. Now it's time to add environment variables, so open the project in the browser

railway open

Add Environment Variables

Back in your Railway dashboard, click on your project and you should see your newly deployed services: Postgres, Server, and Client.

Now you're going to pass each service the correct environment variables. To do this, you first need to tell Railway to generate public domains for client and server.

Go to the server instance's Settings tab, and click Generate Domain. Do the same under the client's Settings.

The Postgres database is already initialized with a domain, so click on the Postgres instance, go to the Connect tab and copy the Postgres Connection URL.

Go back to your server instance and navigate to its Variables tab. Now add the copied Postgres URL as DATABASE_URL, as well as the client's domain as WASP_WEB_CLIENT_URL.

Next, copy the server's domain, move over to the client's Variables tab and add the generated server domain as a new variable called REACT_APP_API_URL.

Having trouble finding these settings?

Postgres Connection URL

Env Variables

And now you should be deployed! 🐝 🚂 🚀

Updates & Redeploying

When you make updates and need to redeploy, just follow steps 3-7 above. Remember, you can connect or disconnect your app to any project in your Railway account by using railway link or railway unlink from within the app's directory.

Customizing the Dockerfile

By default, Wasp will generate a multi-stage Dockerfile that is capable of building an image with your Wasp-generated server code and running it, along with any pending migrations, as in the deployment scenario above. If you need to customize this Dockerfile, you may do so by adding a Dockerfile to your project root directory. If present, Wasp will append the contents of this file to the bottom of our default Dockerfile.

Since the last definition in a Dockerfile wins, you can override or continue from any existing build stages. You could also choose not to use any of our build stages and have your own custom Dockerfile used as-is. A few notes are in order:

• if you override an intermediate build stage, no later build stages will be used unless you reproduce them below
• the contents of the Dockerfile are dynamic, based on the features you use, and may change in future releases as well, so please verify the contents have not changed from time to time
• be sure to supply an ENTRYPOINT in your final build stage or it will not have any effect

To see what your project's (potentially combined) Dockerfile will look like, run: wasp dockerfile

Here are the official docker docs on multi-stage builds. Please join our Discord if you have any questions, or if the customization hook provided here is not sufficient for your needs!