gRPC Streaming - Handling buffer overrun

Question

What will happened in a gRPC streaming service, with a faster producer (server) than the consumer (client)?

• Can the message queue dynamically grow in size?
• If the buffer becomes full, are there mechanisms in place to handle message dropping? If so, how are the dropped messages selected (oldest, newest, random)?

Any insights or best practices for handling buffer overrun in gRPC streaming would be appreciated. Thank you!