Adding an hmget test with more fields and smaller values (so that it becomes listpack)

[slice4e]

Runs memtier_benchmark, for a keyspace length of 1M keys pre-loading
listpack-encoded HASHes with 100 fields and 32 Byte values. After
pre-loading, issues HMGET requesting all 100 fields. Tests HMGET
performance on listpack-encoded hashes where N == L.

Adding this test, in relation to this PR: redis/redis#14903

---

Note

Low Risk
Low risk: adds new benchmark suite YAMLs only, with no production code changes; main risk is increased benchmark runtime/resource usage or misconfigured workloads.

Overview
Adds a new memtier benchmark suite to measure HMGET on listpack-encoded hashes by preloading 1M keys with 100 fields of 32B values and reading all 100 fields with --pipeline 1.

Adds two session-caching playbook suites that preload 100k session:* hashes and apply hash-field expiration via HPEXPIRE, then benchmark read performance for HGETALL and a sibling HKEYS variant designed to highlight CPU-side optimizations under many expired fields.

^{Written by Cursor Bugbot for commit 8954c76. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

[cursor]

Test name missing "playbook-" prefix, mismatches filename

Medium Severity

The name field is memtier_benchmark-session-caching-hash-... but the filename includes playbook- (memtier_benchmark-playbook-session-caching-hash-...). Every other playbook file in the repo has the name field match the filename exactly (minus .yml), including the playbook- segment. Since name is used as the test_name identifier in stats, comparisons, and Parca agent labels, this mismatch breaks the established naming convention and could cause confusion in benchmark reporting and cross-referencing.

Additional Locations (1)

• redis_benchmarks_specification/test-suites/memtier_benchmark-playbook-session-caching-hash-100k-sessions-with-hfe-expired-fields.yml#L1-L2