Page speed results and speed results are really very slow in new update

Being On 4 cores of Intel CPU there is significant drop of speed after i updated the discourse.
I recently Updated my discourse (its pgsql database was also updated) and now its very very slow.
8GB Digital Ocean Droplet.
Results Before this update were - Mobile 75/100

Desktop 85/100
fast-desk

Results Now after this new PGSQL included Update - Mobile 50/100
slow-phone
**Desktop **

I do not have any extra javascript in the page, when i tested, No Alexa, Webmaster etc scripts.
Its a serious issue in Google SEO.
Discourse Team Can you please release a light version with less features for Better speed and Performance.
Or
Please guide me how can i improve upon it.

I also feels the speed difference when i Browse My own site old/new version in incognito mode (Not tested with google.)

I don’t think it would be a good idea to nuke out functionality of discourse in the name of Google pagespeed insights. In my personal opinion, pagespeed insights don’t really focus upon SEO but they rather focus upon the UX of a website.

First load may be a problem (only because all the resources are being loaded from server instead of cache) but once that’s done, there shouldn’t be any performance penalty.

Note that in your tests, Google wasn’t even able to provide an accurate metric of real-world performance… and only looked at possible optimizations. Those possible optimizations don’t seem to take into consideration that Discourse is a JS app.

I think the team is always looking for ways to make Discourse faster… but there’s little evidence that this is a serious SEO issue. Here’s some information from Google on their 2018 Page Speed Update, which is probably the most significant impact Page Speed Score has ever made on ranking:

The “Speed Update,” as we’re calling it, will only affect pages that deliver the slowest experience to users and will only affect a small percentage of queries. It applies the same standard to all pages, regardless of the technology used to build the page. The intent of the search query is still a very strong signal, so a slow page may still rank highly if it has great, relevant content.

So on the one hand, they don’t care that we’re a JS app and apply the same standard to everyone (potentially something to look out for)… but on the other hand this change only impacts a small percentage of queries and content relevance is still the primary metric for ranking.

I also feels the speed difference when i Browse My own site old/new version in incognito mode

Are you able to provide any additional details? Which version were you on previously? Did you use any of the Network tools in the Chrome inspector to measure load speed?

I did run this a few months back and am not seeing anything new here https://developers.google.com/speed/pagespeed/insights/?url=https%3A%2F%2Fmeta.discourse.org&tab=desktop

None of the recommendations are really relevant to us, sure we can shave off a few bytes from avatars and some tiny js scripts, but there is nothing substantial here.

Меня меньше волнует оценка скорости загрузки страницы, а больше — результаты отчёта о пользовательском опыте в Chrome… У меня наблюдаются те же проблемы… Буду рад, если кто-нибудь даст советы по ускорению реального пользовательского опыта, поскольку именно это побуждает посетителей взаимодействовать с вашим сайтом (быстрая загрузка), а не нажимать кнопку «Назад» (медленная загрузка)… Такие отказы отправляют сигнал Google о том, что люди не находят на сайте то, что искали, и именно это приводит к падению позиций в выдаче, а не сама оценка скорости загрузки страницы. Редакция: Кажется, Google действительно заявляет, что использует скорость как фактор ранжирования, но не совсем понятно, какой именно показатель лежит в основе этого.

Похоже, этот сайт использует Fastly в качестве CDN, но всё ещё имеет оценку скорости «медленно» в отчёте Chrome User Experience Report. Возможно, даже с CDN невозможно достичь оценки «умеренно» или «быстро», учитывая характер большого объёма динамического контента…