Deployments

Service Deployments are attempts to build and deliver your application.

Screenshot of Deploy View

All deployments will appear in the deployments view on your selected service. Clicking on the deployment will bring up the build and deploy logs. Each deploy has the option to configure a a Railway provided domain as well as attaching a custom domain.

Deployment Lifecycle

Deployments can be in any of the following states:

  • Initializing
  • Building
  • Deploying
  • Success
  • Failed
  • Crashed
  • Removed
  • Removing

Every Deployment in Railway begins as Initializing - once it has been accepted into our build queue, the status will change to Building.

While a Deployment is building, Railway will attempt to create a deployable Docker image containing your code and configuration (see Builds for more details).

Once the build succeeds, Railway will attempt to deploy your image and the Deployment's status becomes Deploying. If a healthcheck is configured, Railway will wait for it to succeed before proceeding to the next step.

If an error occurs during the build or deploy process, the Deployment will stop and the status will become Failed.

Once the Deployment is live and running, the status will change to Success. A Deployment will remain in this state unless it crashes, at which point it will become Crashed.

When a new Deployment is triggered, older deploys in a Crashed and Success state are eventually removed - first having their status updated to Removing before they are finally Removed. Deployments may also be removed manually.

Bad Gateway

If you get a Bad Gateway when you attempt to visit the deployment URL, it could be that your PORT variable is misconfigured. Railway needs an explicit port to listen on to expose the application to the internet. You can provide a PORT variable under the Variables page in your project. For more information, see Exposing Your App.

Deploy Triggers

A new deploy is triggered when the command railway up is executed. GitHub Services that are linked to a GitHub repo automatically deploy when new commits are detected in the connected branch.

Screenshot of GitHub Integration

You can configure additional deployment triggers such as when a new PR is created using the GitHub Trigger's integration.

Start Command

A start command is the process used to run a Deployment's code. For example, a Python project may have a start command of python main.py, or a NodeJS project may have a start command of npm run start.

Railway automatically configures the start command based on the code being deployed. If your code uses a Dockerfile, the start command defaults to the ENTRYPOINT and/or CMD defined in the Dockerfile. Otherwise, the buildpack used to create the image will determine the start command - see Builds for more details.

Start commands may be overridden for advanced use-cases such as deploying multiple projects from a single monorepo.

When specifying a start command, the behavior of the image depends on type of build:

Screenshot of custom start command configuration

Root Directory

The root directory defaults to / but can be changed for various use-cases like monorepo projects. When specified, all build and deploy commands will operate within that root directory. Additionally, files changed outside the root directory will not trigger a new build.

Watch Paths

Watch paths are gitignore-style patterns that can be used to trigger a new deployment based on what file paths have changed. For example, a monorepo might want to only trigger builds if files are changed in the /packages/backend directory. When specified, any changes that don't match the patterns will skip creating a new deployment. Multiple patterns can be combined, one per line.

Note, if a Root Directory is provided, patterns still operate from /. For a root directory of /app, /app/**.js would be used as a pattern to match files in the new root.

Here are a few examples of common use-cases.

Note, negations will only work if you include files in a preceding rule.

Singleton Deploys

By default, Railway maintains only one deploy per service.

Deployment Actions

Rollback

Users can rollback to previous deploys if mistakes were made. A deployment rollback will revert to the previously successful deployment. Both the Docker image and custom variables are restored during the rollback process.

Screenshot of Rollback Menu

To perform a rollback, click the three dots at the end of a previous deployment, you will then be asked to confirm your rollback.

Logs

Railway allows users to see running logs of your application to help with monitoring. Railway displays the last 10,000 lines of logs available for a deployment.

We maintain logs for inactive deployments as well as active. Under the logs pane, you can search within your logs for certain keywords.

Animation of Log Filtering

Delete Deployments

Users can cancel deployments in progress by clicking the three dots at the end of the deployment tab and select Abort deployment. This will cancel the deployment in progress.

If a deployment is completed, you can delete a live deploy by clicking the the three dots at the end of the deployment tab and select Remove. This will remove the deployment and stop any further project usage.

Restart a Crashed Deployment

When a Deployment is Crashed, it is no longer running because the underlying process exited with a non-zero exit code - if your deployment exits successfully (exit code 0), the status will remain Success.

Railway automatically restarts crashed Deployments up to 10 times. After this limit is reached, your deployment status is changed to Crashed and notifying webhooks & emails are sent to the project's members.

You can restart a Crashed Deployment by visiting your project and clicking on the "Restart" button that appears in-line on the Deployment:

Screenshot of Deploy Options

Restarting a crashed Deployment restores the exact image containing the code & configuration of the original build. Once the Deployment is back online, its status will change back to Success.

You can also click within a deployment and using the Command Palette restart a deployment at any state.

Configurable Restart Policy

Within the Service settings, a user is able to configure a restart policy of either Never, Always, or On-Failure with an optional maximum number of restarts.

How come my GitHub PR wont deploy?

Railway will not deploy a PR branch from a user who is not in your team or invited to your project without their associated GitHub account.


Edit this file on GitHub