- thanks for the suggestion
- this is a more complex issue to discuss:
We dogfood our own app, so we use the homepage as a test of our development on Sparkle 3. In this particular instance I had turned off webp image generation, which is one thing pagespeed is complaining about. I just added webp generation again and there’s essentially no difference in pagespeed.
The “speed index” is biased by Google testing from the US and our site being hosted by DigitalOcean in Amsterdam. Yes this is a real problem, but not a Sparkle problem per se. We don’t currently have a solution for this, if this were your site you might solve it by placing a CDN in front of your Sparkle site (say cloudflare).
The LCP metric is new, public as of last week, and we’re still trying to figure out how to address it. I think we’re doing good, a lot of the weight in the LCP is again the download performance for assets (so the aforementioned testing from a different region). There’s a chrome web vitals plugin and a LCP indicator in the performance section of the web dev tools, and they both show the site doing good.