Add fail_on_cache_miss parameter to CI config variables GraphQL resolver

What does this MR do and why?

This MR adds backend support for a fail_on_cache_miss parameter to the CI config variables GraphQL resolver. This allows clients to explicitly request an error when the reactive cache is not ready, instead of receiving a null response.

This is part of the work to improve error handling in the CI variables GraphQL API, split from !207773 to separate backend and frontend changes.

Problem

Currently, when the ciConfigVariables GraphQL field is queried and the reactive cache is not ready, it returns null without any indication of why the data is unavailable. This makes it difficult for clients to distinguish between:

• Cache not ready (temporary state)
• No variables exist (valid empty state)
• An actual error occurred

Solution

Add an optional failOnCacheMiss boolean argument (defaults to false for backward compatibility) to the ciConfigVariables field. When set to true:

• If the cache is not ready, raise a CacheNotReadyError
• The error is caught and converted to a GraphQL error with a clear message
• Clients can handle this error appropriately (e.g., retry, show loading state)

When false (default), behavior remains unchanged - returns null when cache is not ready.

References

• Parent MR: !207773

How to set up and validate locally

1. Start GDK and navigate to GraphiQL (/-/graphql-explorer)

2. Test default behavior (cache miss returns null):

   query {
     project(fullPath: "your-project-path") {
       ciConfigVariables(ref: "main") {
         key
         value
         description
       }
     }
   }

   Expected: Returns null if cache not ready

3. Test with failOnCacheMiss: true:

   query {
     project(fullPath: "your-project-path") {
       ciConfigVariables(ref: "main", failOnCacheMiss: true) {
         key
         value
         description
       }
     }
   }

   Expected: Returns GraphQL error "Failed to retrieve CI/CD variables from cache." if cache not ready

MR acceptance checklist

Evaluate this MR against the MR acceptance checklist. It helps you analyze changes to reduce risks in quality, performance, reliability, security, and maintainability.