feat(bench): RFC 0033 green (comparative) — cold-vs-warm acquisition, .6

[jensholdgaard]

What

RFC 0033's comparative green slice — discharges RFC0033.6 (the measured-tax-collapses scenario) locally and arms the RFC 0031 harness to measure the cache on the headline corpus (run #20).

• rfc0033_6_measured_tax_collapses (querier crate, un-stubbed): a 64-event widened-template audit history through the production ParquetAuditSink, queried end-to-end twice. Cold: registry_bytes_read == the summed audit file bytes on disk, and the miss write-through published. Warm (fresh querier): registry_bytes_read == the artifact's on-disk size exactly (§5 wording), identical answers, and warm × 10 ≤ cold.
• Measured locally: cold = 518,016 B (64 audit files — the same shape as run feat(miner): implement tokenize() — RFC0001.3 (first red → green step) #8's 513,862 B corpus baseline), warm = 9,288 B → ~55.8×, past the 10× gate with margin.
• Harness arm (rfc0031_comparative.rs): each pair's registry component is classified cold (audit fold) vs warm (bytes equal the published artifact's size) via template_map_outcome(), printed at measurement time and in the report block. Doc note: the artifact persists across pairs by design — production caches persist; the first measured pair may already be warm because the L1 picker's rendering-validation query publishes. Informational until the dispatch run; the frozen-gate math is untouched.
• Abstention finding: write_through abstains iff the JSON would be ≥ the folded audit bytes — at fixture scale a single audit Parquet file (~8 KB) already dwarfs the JSON, so abstention is only reachable for near-empty audit streams; .6 is sized at 64 events so the ratio wins decisively, not just publication.

What this unlocks

Dispatch run #20 measures warm acquisition on otel-demo-v8 — the named §7 condition for unfreezing M_L2. The .7 observability slice (lookup-outcome telemetry via the weaver registry) is the last stub.

Pinned contracts

The bench crate's single-pass/byte tests (honest_total_bytes_breaks_down_additively, one_row_materialization_reads_pages_not_whole_chunks, fixture round-trips) are unmodified and green.

Checks run

cargo fmt --all --check; cargo clippy --all-targets --all-features -- -D warnings (workspace); strict rustdoc (-D broken/private-intra-doc-links); cargo nextest run -p ourios-querier -p ourios-bench — 331 passed, 12 skipped (ignored container/S3 arms + the .7 stub), re-verified after rebasing onto the merged #511.

🤖 Generated with Claude Code

https://claude.ai/code/session_01WQY9wfrfRggqSpMLH8Xj3Y

[coderabbitai]

Warning

Review limit reached

@jensholdgaard, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 30 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 2e692bcf-0330-4748-9d21-33d6be28b82d

📥 Commits

Reviewing files that changed from the base of the PR and between 8fd8e1d and e678c6e.

📒 Files selected for processing (3)

• crates/ourios-bench/src/comparative.rs
• crates/ourios-bench/tests/rfc0031_comparative.rs
• crates/ourios-querier/tests/it/rfc0033_cached_template_map.rs

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch rfc0033-green-comparative

---

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.}

[Copilot]

Pull request overview

This PR advances RFC 0033’s cached template-map work by turning scenario RFC0033.6 into a real, locally-runnable end-to-end test (cold vs warm acquisition) and by augmenting the RFC 0031 comparative harness to classify and report whether each query’s template-map acquisition was a cold audit fold or a warm artifact GET.

Changes:

• Un-stub RFC0033.6 in ourios-querier integration tests: generate a 64-event audit stream, run the same body-rendering query twice, and assert cold/warm byte contracts plus the warm × 10 ≤ cold gate.
• Add template_map_outcome() in the ourios-bench comparative harness and print per-pair cold/warm acquisition classification during measurement and in the final report block.
• Update harness documentation/comments to reflect “template-map acquisition” (vs prior “registry derivation”) terminology in the latency-channel description.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
crates/ourios-querier/tests/it/rfc0033_cached_template_map.rs	Implements RFC0033.6 local arm as a real integration test (cold fold publishes; warm reads artifact; ratio gate asserted).
crates/ourios-bench/tests/rfc0031_comparative.rs	Adds acquisition outcome classification/printing for RFC0033.6’s cold-vs-warm channel in the comparative harness.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.