Why is etcd benchmarking tool put benchmark in sequential mode inserting keys increasing by two / skipping every second key?

[tenzens]

I have an etcd cluster with 5 nodes and I wanted to benchmark it using the benchmarking tool.

I used the following command:
./benchmark put --endpoints $ENDPOINTS --sequential-keys --key-size 2 --key-space-size 128 --val-size 1 --rate 0 --total 127

My thinking was that this would insert keys with increasing value from 0 to 127, interpreted as text I should see all ascii characters (including the characters 0, 1, 2 ... and A, B, C and so on. However when I read all keys with
./etcdctl --endpoints $ENDPOINTS get --from-key ""

or when I feed the output through a hex dump with:
./etcdctl --endpoints $ENDPOINTS get --from-key "" | od -An -tx1

I only get every second key. I do not see a key 1, nor a key 3 (and so on). Why is that?
I checked the source and even though I only brushed over it, at https://github.com/etcd-io/etcd/blob/main/tools/benchmark/cmd/put.go on line 108 it ought to be adding the keys in increasing order.

Any help would be appreciated, maybe i have misunderstood something.