Trouble setting up workflow automation tool on cloud platform

Hey everyone,

I’m stuck trying to get a workflow automation tool running on a popular cloud platform. I’m using containers but keep hitting roadblocks. Could use some help!

Here’s what I’ve done so far:

  1. Set up a basic database:
cloudplatform add-ons:new cloudplatform-postgres:starter --version=11 -a my-project
  1. Created a simple Dockerfile:
FROM workflowtool/base
  1. Made a config file (cloudplatform.yaml):
setup:
  config:
    APP_NAME: "my-project"
    PLATFORM: "cloudplatform.com"
    ENV: "prod"
    TIMEZONE: "UTC"
    TOOL_HOST: "${APP_NAME}.${PLATFORM}"
    TOOL_PORT: "${PORT}"
    TOOL_PROTOCOL: "https"
    TOOL_SECRET: "myverysecretkey"
    WEBHOOK_URL: "https://${APP_NAME}.${PLATFORM}/"
    API_BASE: "https://${APP_NAME}.${PLATFORM}/"
    DATABASE: "postgres"
    DB_HOST: "dbserver"
    DB_NAME: "mydb"
    DB_PORT: 5432
    DB_USER: "dbadmin"
    DB_PASS: "dbsecret"

build:
  docker:
    web: Dockerfile

When I try to run it, I get some weird errors about zombies and permissions. Any ideas what I’m doing wrong? Thanks!

yo, i’ve dealt with this stuff before. sounds like ur container setup might be off. check ur Dockerfile - u need more than just the base image. for zombies, try adding a process manager like tini. permissions issues? make sure ur app’s running as the right user. double-check ur env vars in cloudplatform.yaml too. if ur still stuck, dig into those cloudplatform logs for more clues. good luck!

I’ve been in your shoes before, and I can tell you that setting up workflow automation on cloud platforms can be tricky. From what I see, your setup looks mostly correct, but there are a few things that might be causing issues.

First, make sure your Dockerfile is actually building the workflowtool properly. Just using the base image isn’t enough - you’ll need to add your specific configurations and dependencies.

As for the zombie and permission errors, these often stem from how the container is handling processes. Try adding a proper entrypoint to your Dockerfile that uses a process manager like tini. This can help manage zombie processes.

For permissions, double-check that your app is running as the correct user inside the container. Sometimes, the default user doesn’t have the necessary permissions to write to certain directories.

Lastly, ensure your cloudplatform.yaml file is in the correct location and that all environment variables are properly set. The platform needs to be able to read this file to configure your app correctly.

If you’re still stuck after trying these, it might be worth checking the cloudplatform logs for more detailed error messages. They can often point you in the right direction.

Having worked extensively with cloud-based workflow automation, I can offer some insights. Your setup appears sound, but there are potential issues to address.

Firstly, ensure your Dockerfile is comprehensive. Simply using the base image isn’t sufficient; you need to include specific configurations and dependencies for your workflow tool.

The zombie process issue often stems from improper process management within containers. Consider implementing a process manager like supervisord in your Dockerfile to handle this.

Regarding permissions, verify that your application is running with the correct user permissions inside the container. Incorrect permissions can lead to various operational issues.

It’s also crucial to double-check your environment variables in the cloudplatform.yaml file. Ensure they’re correctly set and accessible to your application.

If problems persist, I’d recommend reviewing the cloud platform’s documentation for any specific requirements or known issues with workflow automation tools. Sometimes, platform-specific nuances can cause unexpected behavior.