additional command tests

benchmarks/EXPIRE/dfly_bench/EXPIRE_reproduce.md

@@ -0,0 +1,43 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
+Note: pick an availability zone that has `c6gn.8xlarge` capacity (AWS suggested
+`us-east-1b/c/d/f`; `us-east-1a`/`1e` were unavailable at run time). Server and
+client share a cluster placement group, so both instance types must have capacity
+in the chosen AZ.
+
+### Test run:
+
+```
+dfbench run-commands-dfly \
+  --id maxqps2 --engines dragonfly,redis,valkey \
+  \
+  --command-template "EXPIRE __key__ 3600" --command-name EXPIRE --command-mutates \
+  --preload-template "SET __key__ __data__" \
+  --key-maximum 100000000 \
+  --test-time 300 --warmup-time 10 --trials 1 \
+  \
+  --dfly-bench-threads 32 --dfly-bench-conns 5 --dfly-bench-pipeline 30 \
+  --dfly-bench-data-size 128 --dfly-bench-key-dist U --dfly-bench-qps 0 \
+  \
+  --dragonfly_version s3://df-use1-pub/v2/dragonfly-aarch64.tar.gz \
+  --dragonfly_num_shards 16 --dragonfly_conn_use_incoming_cpu=true \
+  \
+  --redis_io_threads 8 --valkey_io_threads 8
+```
+
+### Expected results:
+
+**EXPIRE**
+
+| Engine    | Throughput (median) | p50      | p99       | p99.9     | Avg Latency |
+| --------- | ------------------- | -------- | --------- | --------- | ----------- |
+| dragonfly | 8.51M ops/s         | 0.452 ms | 2.705 ms  | 9.896 ms  | 0.551 ms    |
+| redis     | 777.5K ops/s        | 5.777 ms | 13.012 ms | 15.763 ms | 6.162 ms    |
+| valkey    | 1.10M ops/s         | 3.997 ms | 7.949 ms  | 11.370 ms | 4.373 ms    |

benchmarks/INCR/dfly_bench/INCR_reproduce.md

@@ -0,0 +1,42 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
+Note: pick an availability zone that has `c6gn.8xlarge` capacity (AWS suggested
+`us-east-1b/c/d/f`; `us-east-1a`/`1e` were unavailable at run time). Server and
+client share a cluster placement group, so both instance types must have capacity
+in the chosen AZ.
+
+### Test run:
+
+```
+dfbench run-commands-dfly \
+  --id maxqps2 --engines dragonfly,redis,valkey \
+  \
+  --command-template "INCR __key__" --command-name INCR --command-mutates \
+  --key-maximum 100000000 \
+  --test-time 300 --warmup-time 10 --trials 1 \
+  \
+  --dfly-bench-threads 32 --dfly-bench-conns 5 --dfly-bench-pipeline 30 \
+  --dfly-bench-data-size 128 --dfly-bench-key-dist U --dfly-bench-qps 0 \
+  \
+  --dragonfly_version s3://df-use1-pub/v2/dragonfly-aarch64.tar.gz \
+  --dragonfly_num_shards 16 --dragonfly_conn_use_incoming_cpu=true \
+  \
+  --redis_io_threads 8 --valkey_io_threads 8
+```
+
+### Expected results:
+
+**INCR**
+
+| Engine    | Throughput (median) | p50      | p99      | p99.9     | Avg Latency |
+| --------- | ------------------- | -------- | -------- | --------- | ----------- |
+| dragonfly | 10.78M ops/s        | 0.413 ms | 0.831 ms | 1.188 ms  | 0.433 ms    |
+| redis     | 1.56M ops/s         | 2.803 ms | 5.972 ms | 9.891 ms  | 3.074 ms    |
+| valkey    | 1.92M ops/s         | 2.317 ms | 7.349 ms | 10.855 ms | 2.495 ms    |

benchmarks/LPOP/dfly_bench/LPOP_reproduce.md

@@ -0,0 +1,44 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
+Note: pick an availability zone that has `c6gn.8xlarge` capacity (AWS suggested
+`us-east-1b/c/d/f`; `us-east-1a`/`1e` were unavailable at run time). Server and
+client share a cluster placement group, so both instance types must have capacity
+in the chosen AZ.
+
+### Test run:
+
+```
+dfbench run-commands-dfly \
+  --id maxqps2 --engines dragonfly,redis,valkey \
+  \
+  --command-template "LPOP __key__" --command-name LPOP --command-drains \
+  --preload-template "LPUSH __key__ __data__" --preload-items 100 \
+  --command-key-maximum 1000000 \
+  --key-maximum 100000000 \
+  --test-time 300 --warmup-time 10 --trials 1 \
+  \
+  --dfly-bench-threads 32 --dfly-bench-conns 5 --dfly-bench-pipeline 30 \
+  --dfly-bench-data-size 128 --dfly-bench-key-dist U --dfly-bench-qps 0 \
+  \
+  --dragonfly_version s3://df-use1-pub/v2/dragonfly-aarch64.tar.gz \
+  --dragonfly_num_shards 16 --dragonfly_conn_use_incoming_cpu=true \
+  \
+  --redis_io_threads 8 --valkey_io_threads 8
+```
+
+### Expected results:
+
+**LPOP**
+
+| Engine    | Throughput (median) | p50      | p99       | p99.9     | Avg Latency |
+| --------- | ------------------- | -------- | --------- | --------- | ----------- |
+| dragonfly | 10.25M ops/s        | 0.432 ms | 0.871 ms  | 3.507 ms  | 0.456 ms    |
+| redis     | 1.33M ops/s         | 1.562 ms | 11.894 ms | 13.783 ms | 3.588 ms    |
+| valkey    | 1.69M ops/s         | 1.535 ms | 10.042 ms | 11.804 ms | 2.822 ms    |

benchmarks/LPUSH/dfly_bench/LPUSH_reproduce.md

@@ -0,0 +1,42 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
+Note: pick an availability zone that has `c6gn.8xlarge` capacity (AWS suggested
+`us-east-1b/c/d/f`; `us-east-1a`/`1e` were unavailable at run time). Server and
+client share a cluster placement group, so both instance types must have capacity
+in the chosen AZ.
+
+### Test run:
+
+```
+dfbench run-commands-dfly \
+  --id maxqps2 --engines dragonfly,redis,valkey \
+  \
+  --command-template "LPUSH __key__ __data__" --command-name LPUSH --command-mutates \
+  --key-maximum 100000000 \
+  --test-time 300 --warmup-time 10 --trials 1 \
+  \
+  --dfly-bench-threads 32 --dfly-bench-conns 5 --dfly-bench-pipeline 30 \
+  --dfly-bench-data-size 128 --dfly-bench-key-dist U --dfly-bench-qps 0 \
+  \
+  --dragonfly_version s3://df-use1-pub/v2/dragonfly-aarch64.tar.gz \
+  --dragonfly_num_shards 16 --dragonfly_conn_use_incoming_cpu=true \
+  \
+  --redis_io_threads 8 --valkey_io_threads 8
+```
+
+### Expected results:
+
+**LPUSH**
+
+| Engine    | Throughput (median) | p50      | p99       | p99.9     | Avg Latency |
+| --------- | ------------------- | -------- | --------- | --------- | ----------- |
+| dragonfly | 6.02M ops/s         | 0.525 ms | 5.972 ms  | 11.560 ms | 0.778 ms    |
+| redis     | 661.9K ops/s        | 7.110 ms | 11.702 ms | 18.645 ms | 7.239 ms    |
+| valkey    | 866.3K ops/s        | 5.416 ms | 11.628 ms | 15.410 ms | 5.528 ms    |

benchmarks/RPOP/dfly_bench/RPOP_reproduce.md

@@ -0,0 +1,44 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
+Note: pick an availability zone that has `c6gn.8xlarge` capacity (AWS suggested
+`us-east-1b/c/d/f`; `us-east-1a`/`1e` were unavailable at run time). Server and
+client share a cluster placement group, so both instance types must have capacity
+in the chosen AZ.
+
+### Test run:
+
+```
+dfbench run-commands-dfly \
+  --id maxqps2 --engines dragonfly,redis,valkey \
+  \
+  --command-template "RPOP __key__" --command-name RPOP --command-mutates \
+  --preload-template "LPUSH __key__ __data__" --preload-items 100 \
+  --command-key-maximum 1000000 \
+  --key-maximum 100000000 \
+  --test-time 300 --warmup-time 10 --trials 1 \
+  \
+  --dfly-bench-threads 32 --dfly-bench-conns 5 --dfly-bench-pipeline 30 \
+  --dfly-bench-data-size 128 --dfly-bench-key-dist U --dfly-bench-qps 0 \
+  \
+  --dragonfly_version s3://df-use1-pub/v2/dragonfly-aarch64.tar.gz \
+  --dragonfly_num_shards 16 --dragonfly_conn_use_incoming_cpu=true \
+  \
+  --redis_io_threads 8 --valkey_io_threads 8
+```
+
+### Expected results:
+
+**RPOP**
+
+| Engine    | Throughput (median) | p50      | p99       | p99.9     | Avg Latency |
+| --------- | ------------------- | -------- | --------- | --------- | ----------- |
+| dragonfly | 10.67M ops/s        | 0.416 ms | 0.792 ms  | 3.282 ms  | 0.437 ms    |
+| redis     | 1.49M ops/s         | 1.550 ms | 11.361 ms | 11.985 ms | 3.217 ms    |
+| valkey    | 1.81M ops/s         | 1.532 ms | 9.645 ms  | 11.570 ms | 2.639 ms    |

benchmarks/RPUSH/dfly_bench/RPUSH_reproduce.md

@@ -0,0 +1,42 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
+Note: pick an availability zone that has `c6gn.8xlarge` capacity (AWS suggested
+`us-east-1b/c/d/f`; `us-east-1a`/`1e` were unavailable at run time). Server and
+client share a cluster placement group, so both instance types must have capacity
+in the chosen AZ.
+
+### Test run:
+
+```
+dfbench run-commands-dfly \
+  --id maxqps2 --engines dragonfly,redis,valkey \
+  \
+  --command-template "RPUSH __key__ __data__" --command-name RPUSH --command-mutates \
+  --key-maximum 100000000 \
+  --test-time 300 --warmup-time 10 --trials 1 \
+  \
+  --dfly-bench-threads 32 --dfly-bench-conns 5 --dfly-bench-pipeline 30 \
+  --dfly-bench-data-size 128 --dfly-bench-key-dist U --dfly-bench-qps 0 \
+  \
+  --dragonfly_version s3://df-use1-pub/v2/dragonfly-aarch64.tar.gz \
+  --dragonfly_num_shards 16 --dragonfly_conn_use_incoming_cpu=true \
+  \
+  --redis_io_threads 8 --valkey_io_threads 8
+```
+
+### Expected results:
+
+**RPUSH**
+
+| Engine    | Throughput (median) | p50      | p99       | p99.9     | Avg Latency |
+| --------- | ------------------- | -------- | --------- | --------- | ----------- |
+| dragonfly | 5.99M ops/s         | 0.531 ms | 5.922 ms  | 11.373 ms | 0.782 ms    |
+| redis     | 698.8K ops/s        | 6.573 ms | 11.298 ms | 17.824 ms | 6.856 ms    |
+| valkey    | 893.5K ops/s        | 5.197 ms | 11.570 ms | 15.476 ms | 5.359 ms    |

benchmarks/TTL/dfly_bench/TTL_reproduce.md

@@ -0,0 +1,43 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
+Note: pick an availability zone that has `c6gn.8xlarge` capacity (AWS suggested
+`us-east-1b/c/d/f`; `us-east-1a`/`1e` were unavailable at run time). Server and
+client share a cluster placement group, so both instance types must have capacity
+in the chosen AZ.
+
+### Test run:
+
+```
+dfbench run-commands-dfly \
+  --id maxqps2 --engines dragonfly,redis,valkey \
+  \
+  --command-template "TTL __key__" --command-name TTL \
+  --preload-template "SET __key__ __data__" \
+  --key-maximum 100000000 \
+  --test-time 60 --warmup-time 10 --trials 1 \
+  \
+  --dfly-bench-threads 32 --dfly-bench-conns 5 --dfly-bench-pipeline 30 \
+  --dfly-bench-data-size 128 --dfly-bench-key-dist U --dfly-bench-qps 0 \
+  \
+  --dragonfly_version s3://df-use1-pub/v2/dragonfly-aarch64.tar.gz \
+  --dragonfly_num_shards 16 --dragonfly_conn_use_incoming_cpu=true \
+  \
+  --redis_io_threads 8 --valkey_io_threads 8
+```
+
+### Expected results:
+
+**TTL**
+
+| Engine    | Throughput (median) | p50      | p99      | p99.9    | Avg Latency |
+| --------- | ------------------- | -------- | -------- | -------- | ----------- |
+| dragonfly | 10.63M ops/s        | 0.411 ms | 0.851 ms | 1.076 ms | 0.439 ms    |
+| redis     | 1.94M ops/s         | 2.325 ms | 2.980 ms | 3.000 ms | 2.462 ms    |
+| valkey    | 2.33M ops/s         | 2.225 ms | 2.495 ms | 2.499 ms | 2.049 ms    |