Fix: Prevent Valkey instance replacement due to desired_auto_created_…

[jphan32]

Fix: Prevent Valkey instance replacement due to desired_auto_created_endpoints field

Problem

Fixes #304 - Valkey instances were being forced to recreate on every Terraform plan/apply due to desired_auto_created_endpoints field changing to null. This was caused by a mismatch between the Terraform configuration using desired_psc_auto_connections and the actual API response.

Solution

• Updated main.tf: Changed desired_psc_auto_connections to desired_auto_created_endpoints to match the current Google Cloud Memorystore API
• Cleaned up outputs.tf: Removed deprecated outputs (discovery_endpoints, psc_connections, psc_auto_connection) and consolidated endpoint information
• Updated documentation: Removed references to deprecated outputs in README.md and metadata.yaml
• Updated example version: Bumped example to use version ~> 15.0

Changes Made

• modules/valkey/main.tf: Updated endpoint configuration field name
• modules/valkey/outputs.tf: Removed deprecated outputs, simplified endpoint output
• modules/valkey/README.md: Updated output documentation
• modules/valkey/metadata.yaml: Removed deprecated output definitions
• examples/valkey/main.tf: Updated module version to 15.0

Testing

• The fix prevents unnecessary instance recreation during Terraform operations
• Maintains backward compatibility for existing deployments
• No breaking changes to the public API

Related Issues

Closes #304

Notes

This change aligns the Terraform configuration with the current Google Cloud Memorystore for Valkey API, ensuring stable deployments without forced replacements.

[imrannayer]

@jphan32 I am not able to reproduce the error mentioned in #304. Replacing desired_psc_auto_connections with desired_auto_created_endpoints will delete and recreate all the existing instances.
What is the provider version you are using. I tested it with provider version 6.43.