We currently maintain a vendored implementation of TlsChannelImpl for the async driver, which occasionally requires bug fixes and maintenance. An alternative transport implementation, Netty, exists and is fully supported, but requires an additional dependency.

To make informed decisions about the future of TlsChannelImpl, we should start collecting telemetry on which transport layer is being used by clients. If telemetry shows that the majority of users rely on Netty, we can consider deprecating or removing support for TlsChannelImpl in the future.

This data will guide long-term maintenance decisions and help reduce complexity in the async transport layer.