How to improve Vector's consumption capacity?

[loveyang2012]

When using Vector to consume from Kafka with a blackhole sink, a large backlog is observed. How can Vector's throughput be increased? The current configuration file is as follows:

please help me，thinks a lot

The server hardware configuration is 80 cores and 480 GB of memory, and the Vector version is 0.46.