Template Proposal

Inizia esaminando il punto di ingresso della CLI per dokploy deploy e il modo in cui attualmente seleziona le destinazioni del deployment. Definisci il formato di dokploy.yaml e il flag per la selezione dei file, quindi traccia il modo in cui vengono creati o riutilizzati i progetti e i servizi. L’implementazione deve includere il caricamento del file, l’applicazione delle impostazioni dei servizi e il deployment di ogni servizio, con il comportamento documentato per i progetti esistenti e nuovi.

Spinning up server's can be quite a repetitive task especialy when you have multiple servers. As someone using dokploy to manage about 6 projects, I'm proposing creating a serverless-yaml style deployment file that can be used to manage a project.

The idea is to be able to do something like dokploy deploy and it would load dokploy.yaml by default or a user can pass in a flag to specify the deployment file. Then we're able to deploy each service with convig like setting swarm mode, number of replicas, domains etc.

Proposed format

project: uber-clone-project

services:
  - name: geofencer
    image: ghrc.io/geofencer:latest
    replicas: 2
    port: 3000
    domains:
      - hose: geofencer.example.com
      - host: example.com
        path: /v1/geofencer
        tls: true
        redirectToHttps: true
        stripPrefix: true
    env:
      NODE_ENV: production
      API_URL: https://api.example.com
      IP_ADDRESS: ${{project.IP_ADDRESS}}
    resources:
      cpu: 500m
      memory: 512Mi
    configs:
      restart: unless-stopped
      pullPolicy: Always

  - name: backend
    image: dockerub-user/backend:latest
    replicas: 1
    port: 8000
    domains:
      - host: api.example.com
        path: /v1
        tls: true
        stripPrefix: true
    env:
      NODE_ENV: production
      DB_URL: postgres://user:pass@db:5432/mydb
    resources:
      cpu: 300m
      memory: 256Mi
    configs:
      restart: unless-stopped
      pullPolicy: Always

  - name: worker
    image: dockerhub-user/worker:latest
    replicas: 1
    env:
      QUEUE: jobs
      CONCURRENCY: "5"
    resources:
      cpu: 200m
      memory: 256Mi
    configs:
      restart: on-failure

just off the top of my head, if the project doesn't already exist, it should create a new project and use it, if project exists... it should use it.

What are your thoughts?