perf(manager-server): accelerate usage analytics with hourly rollups

[seakee]

Summary

Reuse complete hourly rollups for eligible unfiltered Usage Analytics queries while preserving raw-event fallbacks for filters, percentiles, unsupported timezones, and incomplete checkpoints.
This reduces the 100k-event rollup-owned core from 777 ms / 23.70 MB to 39 ms / 9.51 MB, and publishes a bilingual report for the July 10 performance work.

Scope

• Frontend panel
• Manager Server
• CPA panel mode
• Full Docker mode
• Native packages / release
• Docs / Wiki
• CI / build / tooling

Changes

• Add a shared hourly-rollup reader and migrate Dashboard plus strictly unfiltered Usage Analytics core metrics to it.
• Preserve raw-event parity for edge buckets, filtered or high-dimensional queries, P95/TTFT metrics, DST behavior, and unsupported timezone buckets.
• Add parity, fallback, stability, race, and benchmark coverage plus bilingual performance documentation.

User Impact

Long-window Usage Analytics overview queries use materially less CPU time and memory when the hourly rollup is current. Filtered analytics behavior and API response semantics remain unchanged.

Compatibility / Runtime Notes

• CPA panel mode: No behavior change.
• Manager Server mode: Hourly rollup remains enabled by default and now serves eligible Dashboard and Usage Analytics reads.
• Full Docker / native packages: No deployment change. Set USAGE_DASHBOARD_HOURLY_ROLLUP_ENABLED=false and restart Manager Server to force raw-event reads.

Data / Security Notes

No schema change and no new sensitive-data exposure. Existing rollup tables and sanitized API behavior are unchanged.

Risk / Rollback

Risk level: Medium

Rollback notes:

• Set USAGE_DASHBOARD_HOURLY_ROLLUP_ENABLED=false and restart Manager Server.
• The service will fall back to raw events without deleting existing rollup data.

Verification

• Type check
• Lint
• Tests
• Build
• Manual UI check
• Docs/link check
• Not applicable, docs-only

Commands / evidence:

npm run type-check
npm run lint
npm run test                       # 86 files, 728 tests
npm --workspace apps/web run build:bundle
npm run check:demo-isolation
npm run docs:build
npm run manager-server:test
cd apps/manager-server && go test -race ./...
cd apps/manager-server && go vet ./...
git diff --check origin/main...HEAD

Benchmark evidence:

• Overview: 3.21 s / 34.43 MB -> 2.48 s / 20.24 MB (~23% faster, ~41% lower B/op, ~64% fewer allocations).
• Rollup-owned core: 777 ms / 23.70 MB -> 39 ms / 9.51 MB (~20x faster, ~60% lower B/op, ~92% fewer allocations).
• Final heap profile: approximately 5 MB in use with no retained CPAMP reader/event slice; 200 iterations remained stable near 38-40 ms/op.

Screenshots / Recordings

N/A — backend and documentation changes only.

Docs

• README updated
• Wiki updated
• Release notes needed
• Not needed

Related

N/A