Apollo Server v4 static persisted queries failing after transition from automatic local mode

Transitioning from automatic persisted queries to static persisted queries in Apollo Server v4 causes production 'Query not found' errors. In development, the server successfully caches hashes on-the-fly when the client sends full query strings. However, the production environment is configured to rely on a pre-computed mapping file to restrict arbitrary query execution.

Despite having the client-side build manifest containing the correct hashes, the production server fails to resolve operations when the client sends only the persistedQuery hash. This mismatch occurs even when the server de-serialization store is initialized with the manifest and the Redis cache is properly populated during server startup.

• How can we ensure the Apollo Server de-serialization store correctly synchronizes with the build-time manifest in distributed environments?
• What specific configuration is required to ensure the load balancer does not strip the headers necessary for the hash handshake?