How to improve the write throughput of the dgraph cluster?

What I want to do

when i read docments. i saw the follow content.
Dgraph colocates data per predicate (* P *, in RDF terminology), thus the smallest unit of data is one predicate. To shard the graph, one or many predicates are assigned to a group. Each Alpha node in the cluster serves a single group. Dgraph Zero assigns a group to each Alpha node.
link :

I think this feature will limit write Throughput of the dgraph cluster。
How to improve the write Throughput ?

What I did

i changed this parameter but no effect. currently the write Throughput is only 1800 per second after changed pending_proposals up to 512 or 1024.

--pending_proposals int            Number of pending mutation proposals. Useful for rate limiting. (default 256)

Dgraph metadata

dgraph version

Dgraph version : v20.11.2

@chewxy @JatinDevDG @MichelDiz

Do you have a why you think that?

Add more Alphas and balance the load between them.

What is the maximum mutation per alpha instance? you mean whether increase alphas of the group?

I don’t know. It depends of the context. A machine with i7, 64GB of RAM and NVMe with 6GB/s can respond N times better than a machine with i5, 16GB and a simple SSD. There’s no way to tell you the exact number. But Multiple Alphas(horizontal scaling) helps you to increase the throughput.

Mean, there’s no maximum. Your context that will limit it.

Alphas should be even with in a Cluster.