Skip to content

feat: redesign landing page + batch benchmarks#1

Open
kitsune-hash wants to merge 5 commits intomainfrom
feature/site-improvements
Open

feat: redesign landing page + batch benchmarks#1
kitsune-hash wants to merge 5 commits intomainfrom
feature/site-improvements

Conversation

@kitsune-hash
Copy link
Owner

Changes

Landing Page Redesign

  • Navigation header with anchor links to sections
  • Stats bar (throughput numbers, tech stack)
  • Detailed features grid (9 features with descriptions)
  • How it Works with inline code examples (worker + cURL)
  • Use cases with tags (GPU/CPU, parallel, streaming)
  • Comparison table vs Celery, BullMQ, Temporal
  • Integrations section (Python SDK, REST API, Redis)
  • Step-by-step quickstart guide
  • Why open-source section
  • Professional 4-column footer
  • Improved SEO meta tags

Benchmarks

  • Added batch endpoint section (12.8K-16.5K jobs/s)
  • Added direct Redis bulk benchmark (~83K jobs/s)

Constraints

  • Skeleton.dev classes only, zero custom CSS added
  • SvelteKit/Svelte 5 compatible

Akari added 5 commits February 8, 2026 21:17
- Compare Runqy vs Celery vs BullMQ
- Throughput and P99 latency metrics
- Interactive bar charts
- Transparent methodology section
- Context on when to use Runqy

Benchmark data from reproducible tests:
- 1K, 10K, 50K jobs
- Concurrent submissions (10-100 threads)
- Local Redis, same machine

Key findings:
- BullMQ: highest throughput (native Redis pipelines)
- Celery: good throughput, variable P99 latency
- Runqy: consistent latency, HTTP overhead acceptable for ML workloads
- Add Temporal data (estimated) alongside Runqy, Celery, BullMQ
- Add Temporal analysis card
- Add footnotes about estimated values
- Update meta description
- 1K: 330 workflows/s (was estimated 187)
- 10K: 278 workflows/s (was estimated 157)
- 50K: 240 workflows/s (was estimated 143)

Removed estimate disclaimers.
New throughput: 2921-3001 jobs/s (3x faster than before)
Now beats Celery by 3x, second only to BullMQ.
- Redesigned landing page with proper sections: features, how it works,
  use cases, comparison table, integrations, quickstart, why open-source
- Added navigation header with anchor links
- Added stats bar with throughput numbers
- Added professional 4-column footer
- Added comparison table vs Celery/BullMQ/Temporal
- Added batch endpoint benchmarks (12.8K-16.5K jobs/s)
- Added direct Redis bulk benchmark (~83K jobs/s)
- Improved SEO meta tags
- All styling uses Skeleton.dev classes only (zero custom CSS)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant