Getting started
Deploying on Crowther Cloud
1. Choose your path#
You never pick a build method in the dashboard — the platform looks at your repository and decides. What you choose is how you shape the repository. Find your project below and follow that guide.
2. How your repository is detected#
When a build starts, the platform downloads your branch and checks the root of the repository, in this order. The first match wins:
| If the root contains… | It is built as… | Guide |
|---|---|---|
A Compose file: docker-compose.prod.yml, compose.yaml, docker-compose.yml… | A Container App — every service in the file | Container Apps |
Dockerfile | One image, built from your Dockerfile | Dockerfile apps |
| Neither | One image, built automatically by Railpack from your language and framework | Web services / Static sites |
Only the repository root is looked at
A Dockerfile in backend/ or a Compose file in deploy/ is not found. If your code lives in sub-folders (a monorepo, or frontend/ + backend/), add a Compose file at the root whose build.context points into each folder — see Container Apps → monorepos.
A Compose file that cannot be deployed — typically a development one with bind mounts like ./backend:/app — is skipped in favour of a root Dockerfile if there is one. The reasons are written to the build log. With no Dockerfile to fall back to, the build fails and lists every problem.
3. Your first deploy#
Sign in and connect GitHub
Open cloud.crowtherlabs.com, then New → Web service (or Static site) and click Connect GitHub. This is what lets the platform read private repositories. Public repositories can also be pasted as a URL.
Pick the repository and branch
Choose the repository, then the branch to deploy from the list of its branches. The repository's default branch (often
mainormaster) is picked for you. From now on, every push to this branch deploys automatically; see Automatic deploys.Keep Project on Create a new project for this repository unless you are adding another service built from the same repository: services in one project all build from that project's repository.
Add environment variables
One
KEY=valueper line: database URLs, API keys, secrets. They are encrypted and given to your app when it runs. Read Environment variables before relying on one at build time.Press Deploy and watch the log
You land on the live build log. The build flows straight into the deployment, and the last line tells you where the app is live:
build logDeployment completed — live at https://customer-portal-a1b2c3.cloud.crowtherlabs.com
Running your apps and working together#
- Deployments and builds: workspace history, progress, cancellation, automatic recovery, stop, start and restart.
- Logs and metrics: console output, live following and resource charts.
- Team members: invite people, choose roles and manage access.
- Project settings: the repository, branch and name shared by your services.
- Background workers and schedulers: run them in a Container App alongside your API.
4. Shipping changes#
Pushing to your branch deploys automatically
Push to the service's branch and the new version builds and goes live by itself, usually within a few minutes. Add [skip deploy] to a commit message to push without deploying. How automatic deploys work.
| You changed… | Do this | What happens |
|---|---|---|
| Code | Push to the branch. That's all. | Builds the commit you pushed and deploys it. You get an email if it fails. |
| Environment variables only | Deployments → Redeploy | Restarts the current image with the new values. No rebuild, so it's faster. |
A build-time value (e.g. VITE_API_URL), the branch, or build commands | Settings → Redeploy | Builds the latest commit again with the new settings, then deploys it. |
Building when nothing that affects the image has changed (same commit, same build inputs) reuses the previous image instead of building it again.
5. What every app gets#
- An HTTPS address of the form
https://<name>-<id>.cloud.crowtherlabs.com, working from the first request. Plain HTTP redirects to HTTPS. More on domains. - Automatic restarts if the process crashes.
- Log rotation — each container keeps its last ~30 MB of output.
- A private network shared by the services of the same project.
- Nightly backups of Container App volumes at 02:00 UTC.
Single services have no persistent disk
A Web service, Static site or Dockerfile app starts from a clean filesystem on every deploy. Anything written to disk — uploads, SQLite files, generated reports — is gone after the next deploy. Store files in a database or object storage, or deploy as a Container App with a named volume.
6. Platform facts at a glance#
| Value | |
|---|---|
| Port a Web service must listen on | 3000 |
| Port a Static site is served on | 80 |
| Build time limit | 45 minutes |
| Image architecture | linux/amd64 |
| Container App start-up limit | 15 minutes |
| Replicas per Web service | 1 – 50 |
| Volume backups | Nightly at 02:00 UTC, kept 7 days (Sunday's: 21 days) |
7. Databases#
Run your database inside a Container App, next to the app that uses it, with a named volume for its data. That volume is backed up every night. The Container Apps guide has complete examples for MongoDB, PostgreSQL and Redis.
Note
The PostgreSQL and MongoDB options on the New screen are not ready for use yet. Use the Container App approach until they are.