Environment variable collections

[wertlex]

This PR introduces a new resource called railway_variable_collection.

The railway_variable_collection is essentially a group of environment variables that share a common lifecycle. It can be thought of as multiple railway_variable instances managed together.

The rationale for introducing this new resource is well described in Issue #22. In summary:

• Handling environment variables individually quickly leads to exceeding the Railway API quota. The current implementation can provision any reasonable number of environment variables in a single collection with a fixed number of API calls.
• The current implementation of environment variables does not trigger service redeployment. As a result, even though new environment variables are provisioned, the service continues to consume the previous values. While it was possible to add the required triggers to railway_variable, this approach became problematic for services with 10 or more environment variables.

This new resource addresses these issues by allowing efficient and effective management of environment variables in bulk.

Closes: #22

[wertlex]

Hey! Just asking, if there is something required from my end regarding this PR?

P.S. I've been using these changes in a private fork for couple weeks - seems to be good so far

[wertlex]

@pksunkara thank you for merging it!

Just out of my curiosity, what is the motivation behind going with an array of objects instead of just an object?

[pksunkara]

To make the variable values sensitive since they are secrets.