Add XREADGROUP+XADD benchmarks with COUNT 100 variations (#356)

* Add XREADGROUP+XADD benchmarks with COUNT 100 variations

- Added stream benchmark with XREADGROUP COUNT 100 CLAIM 100 for message claiming
- Added stream benchmark with standard XREADGROUP COUNT 100
- Both use 70-30 XADD-XREADGROUP ratio with 1K consumers
- Pre-loaded with 500K stream entries for concurrent processing tests

* Bumping version from 0.2.60 to 0.2.61

* Refactor stream benchmarks to use 100:1 XADD:XREADGROUP ratio

- Rename benchmark files from 70-30 to 100-1 ratio naming convention
- Update command ratios from 70% XADD / 30% XREADGROUP to 100:1 ratio
- Switch from fixed operation count (-n 1000) to time-based testing (--test-time 120)
- Adjust descriptions to reflect new ratio and expected stream size calculations

The 100:1 ratio ensures the stream always contains elements during benchmark
execution, preventing consumer starvation and providing more consistent
performance measurements. The previous 70-30 ratio could lead to periods
where consumers had no messages to process, affecting benchmark reliability.

* Improve stream benchmark descriptions for clarity

- Add reference to "COUNT 100" so it is clear that we're reading 100 entries for each XREADGROUP command

Author: paulorsousa

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "redis-benchmarks-specification"
-version = "0.2.60"
+version = "0.2.61"
 description = "The Redis benchmarks specification describes the cross-language/tools requirements and expectations to foster performance and observability standards around redis related technologies. Members from both industry and academia, including organizations and individuals are encouraged to contribute."
 authors = ["filipecosta90 <filipecosta.90@gmail.com>","Redis Performance Group <performance@redis.com>"]
 readme = "Readme.md"

redis_benchmarks_specification/test-suites/memtier_benchmark-stream-concurrent-xadd-xreadgroup-100-1-1k-consumers-claim-count100.yml

@@ -0,0 +1,49 @@
+version: 0.4
+name: memtier_benchmark-stream-concurrent-xadd-xreadgroup-100-1-1k-consumers-claim-count100
+description:
+  Starting with a pre-loaded stream of 500K entries, the benchmark tests concurrent stream operations with a consumer group for distributed processing.
+  For each XREADGROUP command, 100 XADD commands are executed.
+  500K initial entries plus 100 * "XADD" - 1 * "XREADGROUP COUNT 100" ≈ 500K entries left in the stream at the end.
+dbconfig:
+  configuration-parameters:
+    save: '""'
+  check:
+    keyspacelen: 1
+  preload_tool:
+    run_image: redislabs/memtier_benchmark:edge
+    tool: memtier_benchmark
+    arguments: '"--data-size" "100" --command "XADD stream-key * field __data__" --command-key-pattern="P" -n 2500 -c 50 -t 4 --hide-histogram'
+  init_commands:
+    - XGROUP CREATE stream-key producer-consumer-group 0 MKSTREAM
+  resources:
+    requests:
+      memory: 4g
+  dataset_name: stream-concurrent-producer-consumer-500k-seed
+  dataset_description: This dataset starts with 500K pre-loaded stream entries, then tests concurrent stream operations with a consumer group for distributed processing.
+tested-commands:
+  - xadd
+  - xreadgroup
+redis-topologies:
+  - oss-standalone
+build-variants:
+  - gcc:15.2.0-amd64-debian-bookworm-default
+  - gcc:15.2.0-arm64-debian-bookworm-default
+  - dockerhub
+clientconfig:
+  run_image: redislabs/memtier_benchmark:edge
+  tool: memtier_benchmark
+  arguments: >
+    --data-size 100
+    --command="XADD stream-key * field __data__" --command-key-pattern="P" --command-ratio=100
+    --command="XREADGROUP GROUP producer-consumer-group __key__ COUNT 100 CLAIM 100 STREAMS stream-key >" --command-key-pattern="P" --command-ratio=1
+    --hide-histogram --key-minimum 1 --key-maximum 1000
+    --test-time 120
+    -c 50
+    -t 4
+  resources:
+    requests:
+      cpus: "4"
+      memory: 2g
+tested-groups:
+  - stream
+priority: 95

redis_benchmarks_specification/test-suites/memtier_benchmark-stream-concurrent-xadd-xreadgroup-100-1-1k-consumers-count100.yml

@@ -0,0 +1,49 @@
+version: 0.4
+name: memtier_benchmark-stream-concurrent-xadd-xreadgroup-100-1-1k-consumers-count100
+description:
+  Starting with a pre-loaded stream of 500K entries, the benchmark tests concurrent stream operations with a consumer group for distributed processing.
+  For each XREADGROUP command, 100 XADD commands are executed.
+  500K initial entries plus 100 * "XADD" - 1 * "XREADGROUP COUNT 100" ≈ 500K entries left in the stream at the end.
+dbconfig:
+  configuration-parameters:
+    save: '""'
+  check:
+    keyspacelen: 1
+  preload_tool:
+    run_image: redislabs/memtier_benchmark:edge
+    tool: memtier_benchmark
+    arguments: '"--data-size" "100" --command "XADD stream-key * field __data__" --command-key-pattern="P" -n 2500 -c 50 -t 4 --hide-histogram'
+  init_commands:
+    - XGROUP CREATE stream-key producer-consumer-group 0 MKSTREAM
+  resources:
+    requests:
+      memory: 4g
+  dataset_name: stream-concurrent-producer-consumer-500k-seed
+  dataset_description: This dataset starts with 500K pre-loaded stream entries, then tests concurrent stream operations with a consumer group for distributed processing.
+tested-commands:
+  - xadd
+  - xreadgroup
+redis-topologies:
+  - oss-standalone
+build-variants:
+  - gcc:15.2.0-amd64-debian-bookworm-default
+  - gcc:15.2.0-arm64-debian-bookworm-default
+  - dockerhub
+clientconfig:
+  run_image: redislabs/memtier_benchmark:edge
+  tool: memtier_benchmark
+  arguments: >
+    --data-size 100
+    --command="XADD stream-key * field __data__" --command-key-pattern="P" --command-ratio=100
+    --command="XREADGROUP GROUP producer-consumer-group __key__ COUNT 100 STREAMS stream-key >" --command-key-pattern="P" --command-ratio=1
+    --hide-histogram --key-minimum 1 --key-maximum 1000
+    --test-time 120
+    -c 50
+    -t 4
+  resources:
+    requests:
+      cpus: "4"
+      memory: 2g
+tested-groups:
+  - stream
+priority: 95