change worker_polling_time_us to 100000 and default core_number

[thweetkomputer]

Here are some reminders before you submit the pull request

• Add tests for the change
• Document changes
• Reference the link of issue using fixes eloqdb/tx_service#issue_id
• Reference the link of RFC if exists
• Pass ./mtr --suite=mono_main,mono_multi,mono_basic

Summary by CodeRabbit

• Chores

  • CPU core allocation now uses a single automatic calculation, capping allocated cores at 90% of available vCPUs (minimum 1).
  • BRPC worker polling interval increased substantially under a specific module (from 1ms to 100ms).

• New Features

  • Added a configuration option to control max write concurrency per core (default: 1), honored from CLI or INI config.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

Walkthrough

Replaced conditional per-branch core_num logic with a single automatic calculation and increased BRPC worker polling interval from 1,000 µs to 100,000 µs (in core/src/data_substrate.cpp). Added a new EloqStore configuration flag and wiring for per-core max concurrent writes (in eloq_store_config.cpp).

Changes

Cohort / File(s)	Summary
Core worker & BRPC config core/src/data_substrate.cpp	Unified core_num calculation to core_num = max(1u, (NUM_VCPU * 9) / 10) (removed prior branch-specific values/logs). Under ELOQ_MODULE_ENABLED, changed worker_polling_time_us from 1000 to 100000.
EloqStore configuration store_handler/eloq_data_store_service/eloq_store_config.cpp	Added new gflag FLAGS_eloq_store_max_concurrent_writes (uint32, default 1) and a new public config field eloqstore_configs_.max_concurrent_writes; populated from CLI flag when non-default or from INI (store.eloq_store_max_concurrent_writes) as fallback and applied to downstream config paths.

Sequence Diagram(s)

(omitted — changes are configuration and local control-flow adjustments that do not introduce a new multi-component sequential flow)

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• remove extra setting of store worker num #261 — Also modifies how Eloq store worker count is handled in core/src/data_substrate.cpp, touching related worker-count logic and gflag usage.

Poem

🐰 I hopped through flags and cores tonight,
Nine-tenths aligned, a tidy sight,
Polling slowed to steady beats,
Writes queued softly, no one cheats,
🥕 Cheers from this small rabbit sprite.

🚥 Pre-merge checks | ❌ 3

❌ Failed checks (3 warnings)

Check name	Status	Explanation	Resolution
Title check	⚠️ Warning	The PR title 'change worker_polling_time_us to 100000' addresses only one component of the changes; the changeset also includes significant modifications to core_num calculation logic and a new eloq_store_max_concurrent_writes configuration flag.	Update the title to reflect all major changes, such as 'Optimize worker polling time and core count configuration' or provide a more comprehensive description of the complete changeset.
Description check	⚠️ Warning	The PR description only contains the template checklist without any substantive content explaining the changes, rationale, testing, or issue references required by the template.	Complete the description by: checking relevant boxes, explaining the changes made, referencing the issue or RFC, documenting test results, and clarifying why these configuration changes were necessary.
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	You can run @coderabbitai generate docstrings to improve docstring coverage.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch polling_time_zc

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Fix all issues with AI agents

In @core/src/data_substrate.cpp:
- Line 797: You changed the BRPC polling interval by calling
GFLAGS_NAMESPACE::SetCommandLineOption("worker_polling_time_us", "100000");
which needs validation and documentation; run comparative performance tests
(measure request latency including tail percentiles and CPU utilization) against
the default polling value, add a short rationale document explaining why 100ms
suits the ELOQ workload (trade-offs of CPU vs latency), add automated tests that
exercise the code path with both default and 100ms settings (asserting
latency/throughput bounds), and record rpcz/bvar profiling output to verify
behavior; update the PR to include the test additions, the performance results
artifact, the rationale text, and mention the exact flag name
worker_polling_time_us and the SetCommandLineOption call so reviewers can locate
the change.

📜 Review details

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 876056c and 1e0f922.

📒 Files selected for processing (1)

• core/src/data_substrate.cpp

[coderabbitai]

Actionable comments posted: 0

🧹 Nitpick comments (1)

core/src/data_substrate.cpp (1)

724-727: Significant increase in automatic core allocation requires documentation.

The core_num formula changed from conditional 60-70% to a fixed 90% of vCPUs. This is a substantial increase that could affect:

• Memory pressure from additional bthreads
• CPU contention with other system processes
• Resource behavior on systems with data store disabled (previously used 70%, now 90%)

Please document the rationale for this change and verify performance under various deployment scenarios.

📜 Review details

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 4e593b2 and 07b909b.

📒 Files selected for processing (2)

• core/src/data_substrate.cpp
• store_handler/eloq_data_store_service/eloq_store_config.cpp

🧰 Additional context used

🧬 Code graph analysis (1)

store_handler/eloq_data_store_service/eloq_store_config.cpp (2)

core/include/data_substrate.h (1)

• CheckCommandLineFlagIsDefault (313-324)

store_handler/eloq_data_store_service/main.cpp (3)

• CheckCommandLineFlagIsDefault (89-101)
• CheckCommandLineFlagIsDefault (89-89)
• config_reader (163-163)

🔇 Additional comments (2)

store_handler/eloq_data_store_service/eloq_store_config.cpp (2)

127-129: LGTM!

The new flag follows the established pattern and the default value of 1 is appropriately conservative for write concurrency control.

---

515-520: LGTM!

The configuration reading follows the established pattern used throughout this file, correctly prioritizing command-line flags over INI config values.

[zhangh43]

+1