Benchmark metadata and payload cache updates

[mjc]

Summary

• add a divan benchmark for hybrid cache update costs
• separate metadata-only, retained-payload, and mixed update workloads

Branch-added tests

• The benchmark defines isolated unique-key workloads for metadata-only, retained-payload, and mixed updates, with one-sample iterations and explicit hybrid-cache shutdown so the measured paths are comparable and bounded.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI (base), Organization UI (inherited)

Review profile: ASSERTIVE

Plan: Advanced

Run ID: d291a245-7c8a-4aff-8386-fe1295780cf0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

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]

🟡 Changes recommended

The benchmark’s timed region currently includes input construction (formatting/allocation), which undermines the stated goal of isolating cache update costs.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Pull request overview

This PR adds a new Divan benchmark to quantify hybrid cache update costs, splitting the workload into metadata-only, retained-payload, and mixed update scenarios to make comparisons between update paths more explicit and bounded.

Changes:

• Registers a new cache_metadata_payload benchmark target in Cargo.toml.
• Adds benches/cache_metadata_payload.rs implementing three hybrid-cache update workloads (metadata-only, payload-only, mixed) with explicit cache shutdown.

File summaries

File	Description
Cargo.toml	Adds the new benchmark target so it can be executed via cargo bench --bench cache_metadata_payload.
benches/cache_metadata_payload.rs	Implements the new Divan benchmark scenarios for hybrid cache metadata/payload update costs.

Review details

Suppressed comments (2)

benches/cache_metadata_payload.rs:76

• The benchmark currently measures MessageId/response construction (formatting + allocation) along with the cache update. If the goal is update cost, move per-iteration input creation into with_inputs() and benchmark only upsert_ingest.

    bencher.bench(|| {
        let sequence = sequence.fetch_add(1, Ordering::Relaxed);
        runtime.block_on(cache.upsert_ingest(
            message_id(sequence),
            black_box(article_response(sequence)),

benches/cache_metadata_payload.rs:96

• This mixed benchmark times key/response construction and atomic increments along with the cache updates. Using with_inputs() keeps the timed region focused on the cache operations so the result reflects metadata/payload update costs rather than formatting/allocation overhead.

    bencher.bench(|| {
        let sequence = sequence.fetch_add(2, Ordering::Relaxed);
        runtime.block_on(async {
            cache
                .record_backend_has_status(

• Files reviewed: 2/2 changed files
• Comments generated: 1
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.