Updates about our architecture and trial server

[gweng]

Hi all,

I have launched an new trial instance for deploying our services. And I also updated some parts of our architecture:

1. Any connection between arbitrary services without any well-known protocol, like Server-Database via SQL agent, should using RabbitMQ as the most general message passing mechanism

For example, our Analyzer may directly fetch data via the classic database agent in SQL, but it will put commands and notifications to other services in the RabbitMQ for other services.
Another example is if API Agent want to receive/notify there are some new changes of weather or locations, it will also receive the command from and put the result in RabbitMQ.

2. All services should be encapsulated as a Docker container before it can be deployed on the trial server

One exception is if how to encapsulate such service is still unclear, it could be deployed as a interactive container like:

    docker run -i -t ubuntu /bin/bash      

And then you can try anything inside the container as an ordinary VM. However, some differences should be noted. For example, the most critical is committing after doing the change:

    docker commit <container> <some_name>

If you don't do this, all your changes will be lost.

But this is only for those who is closing to deploy the ready service. If you're service is still constructing on your local console, and you're not going to push the whole content to our server, you don't need to do that.
However, it is still encouraged to build and encapsulate such service on your own local console, because this will save the time to migrate to that.

If you have no idea about how to build things in Docker, or you have encountered any difficulty about that, please post a thread here.

3. How to access the trial server

We will soon have a discussion about how to make it accessible to developers, if necessary.

Thanks.