deploy.party is a simple and easy-to-use deployment tool for your applications.
- π Deploy any application with a simple UI
- π Deploy directly from GitLab or GitHub
- π Deploy docker image to local registry or other registries
- π Start databases with a single click (MongoDB, MariaDB)
- π Create custom containers with custom docker-compose and Dockerfile
- π Split your containers into projects
- π Automatic SSL and www-redirect
- π Simple backups to S3 compatible storage
- π Integrated terminal for direct access to the container beta
- π See logs of your containers
- π See statistics of your containers
- Server access with root permissions
- Configure DNS with A record for
YOUR_URLand*.YOUR_URLtoSERVER_IP.
Replace follow placeholders with your values:
SERVER_IP: IP of the server
ssh root@SERVER_IPbash -c "$(curl -fsSL https://raw.githubusercontent.com/lenneTech/deploy.party/main/install.sh)"For automated deployments, you can pass parameters directly to the install script:
# Download install script
curl -fsSL https://raw.githubusercontent.com/lenneTech/deploy.party/main/install.sh -o install.sh
chmod +x install.sh
# Local installation
./install.sh --local --name my-instance --username admin --password secret123
# Production installation
./install.sh --name my-instance --url example.com --email admin@example.com --username admin --password secret123-l, --local- Run deploy.party locally (skips URL and email prompts)-n, --name NAME- Instance name-u, --url URL- Base URL (required for production setup)-e, --email EMAIL- Email for SSL certificates (required for production setup)--username USER- Username for Traefik dashboard--password PASS- Password for Traefik dashboard-h, --help- Show help message
# Minimal local setup
./install.sh -l -n dev-server --username admin --password mypassword
# Full production setup
./install.sh -n production \
-u mydomain.com \
-e admin@mydomain.com \
--username traefik-admin \
--password secure-password-123
# Get help
./install.sh --helpNote: Any missing parameters will be prompted interactively during installation.
https://www.digitalocean.com/community/tutorials/how-to-protect-ssh-with-fail2ban-on-ubuntu-22-04
nano /var/opt/deploy-party/data/.envAfter you have configured the .env file, you can reconfigure the deploy.party with the following command:
sh /var/opt/deploy-party/reconfigure.shsh /var/opt/deploy-party/update.shOlder installs started the local registry with -p 5000:5000. It has no authentication, and Docker bypasses UFW, so every image (including its .env) was readable from the internet unless the provider firewall blocked port 5000. New installs bind it to 127.0.0.1. For an existing install, run once:
cd /var/opt/deploy-party
curl -fsSL https://raw.githubusercontent.com/lenneTech/deploy.party/main/secure-registry.sh -o secure-registry.sh
bash secure-registry.sh --check # exit 2: still public
bash secure-registry.sh # switch; keeps the data volume, rolls back by itself on any difference
bash secure-registry.sh --cleanup # after the next successful deploy: remove the stopped old containerThe script refuses on a multi-node swarm (other nodes pull from this registry) and while a push is running. --rollback restores the previous container. Once secure-registry.sh sits next to update.sh, every update runs it as well.
A build used to take the Dockerfile and docker-compose.yml that were on disk. Only a deploy from the UI and a tag build wrote them again from the container settings, so a changed setting went live only after a deploy, and the healthcheck of NODE containers and the rollback settings of CUSTOM containers reached existing containers only that way. With NSC__REFRESH_BUILD_FILES=true every build writes both files from the settings first; a file whose content changes is kept as Dockerfile.prev or docker-compose.yml.prev, and the build log says which file changed. New installs set it. For an existing install, check first what the first builds would change:
docker exec $(docker ps -qf name=deploy-party_api) node dist/src/server/scripts/build-files-drift.jsOne line per container: UNCHANGED, EXPECTED (only the healthcheck, the update and rollback settings or a basic-auth label in the new hash format) or OTHER with the keys that differ (never values). Settle every OTHER first: settings that were changed but never deployed, or a file edited by hand on the server. Then add NSC__REFRESH_BUILD_FILES=true to /var/opt/deploy-party/.env and run update.sh.
POST /webhook used to accept every request: whoever knew a GitLab project id could start builds, and for containers with a tag pattern choose the tag that gets deployed. Every source now has its own secret (created by a migration for existing sources), which GitLab sends as X-Gitlab-Token; new hooks created from the UI carry it. NSC__WEBHOOK_TOKEN_MODE switches the check: off (default), report (accept, but log every request without the right secret) or enforce (build only with the secret; otherwise HTTP 401). New installs set enforce. For an existing install:
- Run
update.sh(the migration creates the secrets), setNSC__WEBHOOK_TOKEN_MODE=reportand runupdate.shagain. - Put the secrets into the existing hooks. The GitLab token goes in on stdin and needs Maintainer rights with scope
api; the script prints only project and hook ids:printf '%s\n' "$GITLAB_ADMIN_TOKEN" | docker exec -i $(docker ps -qf name=deploy-party_api) \ node dist/src/server/scripts/webhook-secret-hooks.js --gitlab-host gitlab.example.com # dry run printf '%s\n' "$GITLAB_ADMIN_TOKEN" | docker exec -i $(docker ps -qf name=deploy-party_api) \ node dist/src/server/scripts/webhook-secret-hooks.js --gitlab-host gitlab.example.com --apply
- Watch the API log for
came with no X-Gitlab-Token/a wrong X-Gitlab-Token(for example CI jobs that call the webhook themselves). When nothing shows up any more, setNSC__WEBHOOK_TOKEN_MODE=enforceand runupdate.sh.
The API served Swagger UI under api.YOUR_URL/api and the OpenAPI document under /api-json to everyone. Both are off now; deploy.party itself does not use them. To switch them on for an instance, set NSC__SWAGGER=true in the .env and run update.sh.
Now Traefik UI will be available at lb.YOUR_URL and deploy.party at YOUR_URL.
Traefik has basic auth lock and can be accessed with the USERNAME and PASSWORD you set in the init.sh script.
On local setup:
- Deploy Party: http://[IP]:3001
- Minio UI: http://[IP]:3004
- Deploy Party API: http://[IP]:3000
- Deploy Party Terminal API: http://[IP]:3002
- Minio API: http://[IP]:3003
Default credentails are:
admin@deploy.party
deploy-party
Please create a new admin account and delete the old one.
- Click on plus in topbar and create new member
- Fill form and choose the right role
- The invited user receives an e-mail with the request to set a password.
- Go to sources
- Click on plus in topbar and create new source
- Enter name of source for example GitLab
- Enter the URL of the GitLab instance for example https://gitlab.com
- Enter an access token for the GitLab API GitLab Docs
- Use a project or group access token, not a personal one. deploy.party uses the token to clone, to list projects, branches and releases, and to create the webhook, so it needs the role Maintainer and the scope
api. Addread_registry, write_registryonly if you push to the GitLab registry. - A personal access token reaches every project of its owner. If it leaks, all of them are exposed, not only the deployed ones. A project token (
/GROUP/PROJECT/-/settings/access_tokens) or group token (/groups/GROUP/-/settings/access_tokens) is limited to what deploy.party deploys. - Give the token an expiry date and note where it is used.
- Since DEV-4027 the token is passed to
git cloneas a header and no longer stored in the clone (.git/config), in the image or in the build log.
- Use a project or group access token, not a personal one. deploy.party uses the token to clone, to list projects, branches and releases, and to create the webhook, so it needs the role Maintainer and the scope
- Click on plus in topbar and create new project
- Enter name for project
- Click on plus in topbar and create new container
- Select the created project
- Enter the name of the container
- Choose the kind of container (
APPLICATION,DATABASE,SERVICE,CUSTOM)APPLICATION: A normal applicationDATABASE: A databaseSERVICE: A service like a directus, ...CUSTOM: A custom container with custom docker-compose and Dockerfile
- Select the source
- Enter the image name
- Select registry
- Enter commands
- Enter environment variables
- Enter ports
- Start the container
docker stack rm traefikWait if traefik container not shown under docker ps before you run the next command.
cd /var/opt/deploy-partyexport $(cat .env | grep -v '#' | awk '/=/ {print $1}') && docker stack deploy -c docker-compose.traefik.yml traefikcd /var/opt/deploy-partysh reconfigure.shAs of the latest version, deploy.party uses Traefik v3 by default. If you have an existing installation running Traefik v2, please see the comprehensive migration guide:
π Complete Migration Guide (MIGRATION.md)
The guide includes:
- β Automated migration script (recommended)
- β Manual migration steps
- β Rollback instructions
- β Troubleshooting guide
- β Post-migration checklist
For a quick automated migration:
cd /var/opt/deploy-party
curl -fsSL https://raw.githubusercontent.com/lenneTech/deploy.party/main/upgrade-traefik-v3.sh -o upgrade-traefik-v3.sh
chmod +x upgrade-traefik-v3.sh
# Interactive mode (prompts for API URL and token)
./upgrade-traefik-v3.sh
# Or with parameters for automation
./upgrade-traefik-v3.sh --api-url https://api.deploy.party --api-token dp-xxx...
# Skip container migration
./upgrade-traefik-v3.sh --skip-migration
# Skip deploy.party update
./upgrade-traefik-v3.sh --skip-deploy-party-updateWhat you'll need:
- π API URL (optional - auto-detected from APP_URL if available)
- Format:
https://api.YOUR_DOMAINorhttps://YOUR_DOMAIN:3000 - The script will try to auto-detect or prompt you
- Format:
- π API Token (optional but recommended for automatic container migration)
- Get it from: Settings β API Keys in your deploy.party UI
- Starts with
dp- - Can be passed via
--api-tokenflag or entered interactively
What happens:
β οΈ Traefik v3 requires middleware syntax changes (@swarmsuffix)- π All deployed containers will be restarted with updated configuration (rolling update)
- β±οΈ Each container has ~5-10 seconds downtime during restart
- β Docker Compose files are automatically regenerated
- π deploy.party is updated to the latest version