Image fetch and display optimization on the landing page (LCP 404 ms → poster priority, sizing, caching)
Mantenedores costumam responder em até 1 dia
Ninguém assumiu esta issue ainda.
Avaliação
- Dificuldade
- 4/5
- Tempo estimado
- 3-5 dias
- Facilidade para iniciantes
- 35/100
- Tipo de issue
- Refatoração
- Clareza
- Razoavelmente clara
- Status de atividade
- Ativa
- Stack de tecnologia
- html, javascript
- Domínio
- frontend, performance
Direção de pesquisa
Start in website/site/src/components/FilmScreen.astro where the hero video poster and attributes are set, then check Base.astro for a possible preload and the inline module in the built index.html. Caching needs a new website/site/public/_headers file and wrangler.jsonc, and the poster resizing belongs in the pnpm run media pipeline under website/hero-animation/. Done means the acceptance checklist holds: High poster priority in DevTools, a smaller poster on a 1318×741 viewport, Cache-Control on /media/*, and LCP under 300 ms.
Escrita pelo modelo de indexação a partir do texto da issue.
Descrição
Context
LCP analysis of ema.tnsor.network (measured 2026-10-16): LCP is the hero <video>'s poster frame (/media/hero-poster.webp), landing at 404 ms with a 312 ms render delay after the image finishes downloading (~92 ms). Full diagnostics:
| Metric / Attribute | Value | Finding |
|---|---|---|
| LCP Timing | 404 ms | Total render delay is 312 ms. |
| Image Priority | Medium | Failed fetchpriority=high check. |
| Resolution | 1920×1080 | Oversized for the 1318×741 display area. |
| Potential savings | 39.1 KB | Oversized resolution wastes bytes. |
| Cache policy | max-age=0 |
Forces re-validation on every request. |
Render delay breaks down to a ~205 ms layout task plus script execution (line 19, col 5131 — inline module in the built index.html) competing with the poster's first paint.
Current markup (FilmScreen.astro)
<video muted loop playsinline preload="auto" poster={poster} data-autoplay data-low-src={lowSrc} aria-label={title}>
No fetchpriority, no responsive source, and Cloudflare Workers static assets currently emit no cache headers (no _headers file; wrangler.jsonc assets.directory only).
Recommendations
- Elevate resource priority —
fetchpriority="high"on the video (applies to the poster fetch). Optionally add a<link rel="preload" as="image">for the poster inBase.astrowhen the hero is above the fold. - Responsive sizing — the poster is 1920×1080 but the display area measured 1318×741. Serve a smaller default with
srcset/<picture>(e.g. 960/1318/1920 widths), or accept the ~39 KB penalty on desktop. Note the poster doubles as theog:image, so generate both variants from the same source inpnpm run media. - Minimize main-thread work — the ~205 ms layout task + inline script at first paint delay the render even after the bytes arrive. Audit the initial inline module (spark canvas init, IntersectionObserver setup, reveal class) for anything that can yield or move behind
requestIdleCallback. - Improve caching — static assets are content-hashed only in name (they aren't —
hero-poster.webphas no hash). Add awebsite/site/public/_headersfile (Cloudflare Pages syntax works with Workers static assets) mapping/media/*to a longmax-age+immutable, or hash filenames. Careful:hero.mp4changes with each hero re-render, so prefer_headerswith a boundedmax-age(e.g. 1 week) overimmutableunless filenames get hashed.
<!-- Example: adding fetchpriority to a video poster -->
<video poster="/media/hero-poster.webp" fetchpriority="high">
<!-- ... -->
</video>
Note: findings above were identified on a live page in DevTools. Adapt to the stack (Astro components in website/site/src/components/FilmScreen.astro, build pipeline in website/hero-animation/) rather than applying literal CSS overrides.
Acceptance
- Poster fetch reports
Highpriority in DevTools network - Poster transfer ≤ 1318×741-equivalent size on that viewport (or documented decision to keep 1080p for large screens)
-
_headersserving/media/*withCache-Control: public, max-age≥604800(verified withcurl -sI https://ema.tnsor.network/media/hero-poster.webp | grep -i cache) - LCP re-measured with the same methodology; target < 300 ms
- Linguagem predominante
- TypeScript
- Estrelas
- 829
- Forks
- 234
- Merge médio
- 22h 43min
- PRs com merge (30d)
- 52
Preparar o ambiente
- Inclui um Dockerfile ou arquivo Docker Compose
- Tem um modelo de pull request
- Ler o guia de contribuição
Primeiros passos
- Leia a issue inteira e depois o guia de contribuição do projeto.
- Comente na issue dizendo que vai assumir — evita que duas pessoas façam o mesmo trabalho.
- Faça um fork do repositório e trabalhe em uma branch.
- Abra um pull request que referencie o número da issue.
Mais de cameri/nostream
-
Website
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 84/100
Mantenedores costumam responder em até 1 dia
-
Dificuldade 2/5 Meio dia Facilidade para iniciantes 62/100
Mantenedores costumam responder em até 1 dia
-
Add the official Nostream Telegram channel and development chat to the websiteTalvez já em andamento @chappie-daemon assumiu hoje. Abertadocumentation
cameri/nostream#878 · 1 responsável ·
Mantenedores costumam responder em até 1 dia
-
feat(admin): Restyle admin console to match nostream.dev and fix Grafana chart embedsTalvez já em andamento @Ferryx349 assumiu há 1 dia. AbertaAdmin Console enhancement
cameri/nostream#874 · 1 responsável ·
Mantenedores costumam responder em até 1 dia
-
factory rewrite
Dificuldade 5/5 Mais de uma semana Facilidade para iniciantes 35/100
Mantenedores costumam responder em até 1 dia
Todas as issues de cameri/nostream
Issues semelhantes
-
Dificuldade 1/5 Menos de uma hora Facilidade para iniciantes 65/100
521xueweihan/HelloGitHub#4025 ·
-
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 82/100
voidzero-dev/void#478 ·
Mantenedores costumam responder em até 1 dia
-
bug
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 69/100
Small-tailqwq/dsh-deep-whale#207 ·
Mantenedores costumam responder em até 1 dia
-
The inspector never shows the instances field, so scaling out is unreachableTalvez já em andamento @Mrkewl assumiu hoje. Aberta
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 76/100
xevrion/breakscale#96 ·
Mantenedores costumam responder em até 4 dias
-
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 70/100
Flux159/mcp-server-kubernetes#384 ·