You have an application using GCP's Cloud Spanner, and you've noticed that write operations are becoming a bottleneck as your application scales. What is the best approach to resolve this issue?

*Explanation:* The most effective way to address the bottleneck in write operations for your Cloud Spanner-based application is to increase the number of Spanner nodes. This approach can improve write throughput and scalability. Sharding is unnecessary as Cloud Spanner handles sharding automatically. Implementing caching with Cloud Memorystore may help with read operations but won't solve the write bottleneck. Converting to a regional instance or using batch writes may not be the most suitable solutions for this specific issue.