Launch a New App on Fly.io

All new apps now deploy to V2 of the Fly Apps platform, running on Fly Machines. Most docs focus on Apps V2, but we still include information specific to the legacy Fly Apps V1 where appropriate.

We'll be migrating all V1 apps in phases. Learn more about how and why we're getting off Nomad.

You can also migrate your V1 app yourself using our migration tool or manually.

To create a brand new app on Fly.io, run

fly launch

from the root directory of your project.

If you're in a hurry to try your first launch, we also have a condensed Hands-on. You may also be interested in our language- and framework-specific launch guides.

Ingredients of a Ready-to-deploy Fly App

Here are the components of a successful launch, ready for the first deployment:

Framework Launch Scanners

Depending on your project, fly launch may be able to look at your app's source code and get through that list, straight to a deployed Fly App, in one swoop. This is most likely for frameworks on which Fly.io has people specializing full time. Right now that's Elixir/Phoenix, Laravel, Rails, and Django, but we have launch guides for various other languages and frameworks.

Our best scanners furnish a Dockerfile from which your app's image will be built. Some of our terrible older scanners may invoke buildpacks, which tend to be slow and brittle.

Running fly launch in a directory containing a working Django app (it happens to be the one from our Django getting-started example):

fly launch
Creating app in /Users/chris/FlyTests/django
Scanning source code
Detected a Django app
? Choose an app name (leave blank to generate one): 
...

The flyctl Django scanner has taken ownership of my launch. Visit our Django guide to see how that story will end. (Spoiler: it has a happy ending.)

Custom Launch

You can nudge fly launch to better suit your project.

Choose How the Docker Image Is Built

Tell fly launch how you want to get the Docker image for your app, using either the --image or --dockerfile option, or by catching the Dockerfile launch scanner's attention with the presence of a Dockerfile in your app's root directory. The Dockerfile scanner doesn't do a lot of configuration, but it prevents other scanners from taking over.

The actual Docker image build (or image pull) for a Fly App takes place during deployment. fly launch sets the stage by recording how to build, or get, the image, and both the first and all later deploys use that information.

Customize the Configuration File

You can provide your own fly.toml and fly launch will offer to read it in, in lieu of the default config. That sets a starting point for the app configuration, and under some circumstances a framework launch scanner may overwrite parts of it.

There are also a number of other options you can use to exert control over fly launch.

If fly launch doesn't have a scanner that can set up your app automatically, it will still initialize a new Fly App in your Fly.io organization and provide you with a default app configuration that's a reasonable starting point for a simple web app.

It's also possible to perform an entirely manual "launch", skipping all the launch scanners and full-service resource provisioning, using fly apps create, a hand-crafted (or copied) fly.toml, and step-by-step resource provisioning, followed by fly deploy.

An Example

Here's me launching my Flask app that I've written and tested using the local dev server:

fly launch 
Creating app in /Users/chris/FlyTests/hello-gunicorn-flask
Scanning source code
Detected a Python app
Using the following build configuration:
        Builder: paketobuildpacks/builder:base
? Choose an app name (leave blank to generate one): 

If I carry on with this buildpack-based build, this story does not end with a working Flask app. Dockerfile-based deployments are much simpler and faster anyway, so I decide to use a Dockerfile. (I hit ctrl-C to stop the launch.)

Conveniently, there's also already a Dockerfile that works with this app.

With a Dockerfile in my working directory, if I run fly launch again, the Dockerfile launcher takes over before the generic Python buildpack one has a chance to.

fly launch
An existing fly.toml file was found for app testrun
? Would you like to copy its configuration to the new app? No

I have a fly.toml in my working directory from cloning the fly-apps/hello-gunicorn-flask repo to my local machine, but I want to use whatever config fly launch gives me, so I answer No to that question.

Creating app in /Users/chris/FlyTests/hello-gunicorn-flask
Scanning source code
Detected a Dockerfile app

There's the Dockerfile scanner taking over. Now some general app configuration:

? Choose an app name (leave blank to generate one): testrun
? Select Organization: Chris (personal)
Some regions require a paid plan (fra, maa).
See https://fly.io/plans to set up a plan.

? Choose a region for deployment: Toronto, Canada (yyz)

This is the region where flyctl will start Machines for this app if not otherwise specified. The first deployment will put its Machine(s) in that region.

Created app 'testrun' in organization 'personal'
Admin URL: https://fly.io/apps/testrun
Hostname: testrun.fly.dev
? Would you like to set up a Postgresql database now? No
? Would you like to set up an Upstash Redis database now? No

fly launch offers to provision Fly Postgres and Redis by Upstash databases. I don't need them for my very simple web app.

Wrote config file fly.toml
? Would you like to deploy now? Yes
==> Building image
Remote builder fly-builder-crimson-dew-6190 ready
...

It's downloaded the new app's configuration into fly.toml and offered to deploy straight away. I accept. By default, the Docker build happens on one of Fly.io's remote builder Machines.

fly launch always rewrites fly.toml even if it's just writing an identical config back into the file.

Aside: In the case of my Dockerfile app, I ended up with the default fly.toml, configured with an HTTP service suitable for a basic web app, but with one small tweak. My Dockerfile contains the line:

EXPOSE 4999

My Flask app listens on port 4999, not the more common (and default for fly.toml) port 8080.

Fly.io doesn't care about EXPOSE statements in Dockerfiles directly, because we don't actually run containers, and services via our proxy are configured in fly.toml, not in the Dockerfile. But as a convenience, the Dockerfile fly launch scanner catches EXPOSE if it's there, and fills in internal_port on the public HTTP [[services]] section accordingly.

Coming back to the deployment, skipping some of the build output:

...
--> Pushing image done
image: registry.fly.io/testrun:deployment-01GWAPGQWQ5N1HY5651D5DX5B0
image size: 141 MB
Provisioning ips for testrun
  Dedicated ipv6: 2a09:8280:1::5b:297
  Shared ipv4: 66.241.125.144
  Add a dedicated ipv4 with: fly ips allocate-v4

Because I had an HTTP service configured, and no public IP addresses, these were provisioned on deployment.

Process groups have changed. This will:
 * create 1 "app" machine
No machines in group 'app', launching one new machine
  Machine 17811122f5d089 [app] update finished: success
  Finished deploying

The first deployment has finished!

I haven't explicitly configured process groups, so my app gets a single Machine assigned to the default app process. With fly status I can see that this machine is running in yyz (Toronto), where I told fly launch I wanted my app to be deployed.

fly status
App
  Name     = testrun                                        
  Owner    = personal                                   
  Hostname = testrun.fly.dev                                
  Image    = testrun:deployment-01GWAPGQWQ5N1HY5651D5DX5B0  
  Platform = machines                                   

Machines
ID              PROCESS VERSION REGION  STATE   HEALTH CHECKS   LAST UPDATED         
17811122f5d089  app     1       yyz     started                 2023-03-24T20:56:45Z

I can scale out by adding Machines in other regions if I want to get close to users in more corners of the world.

To check that my new web app is actually working, I run

fly open

to visit my app in the browser!

After fly Launch

If you've run fly launch but haven't deployed yet, want to change your configuration, update your app source, or change or provision platform resources such as public IP addresses, Fly Volumes, app secrets, Fly Postgres clusters, or integrated resources like Redis by Upstash, you can go ahead and do so, and (re)deploy with fly deploy.

Once your app is deployed you'll probably want to check up on it using one or more of the techniques in Get Information About an App. When you're ready to go to production, scale up to two or more Machines for high availability.