cache tags being free doesnt change the behaviour where worker responses wont be cached unless they
cache tags being free doesnt change the behaviour where worker responses wont be cached unless they do it themselves with cache api
Ecache tags can be used alongside the cache api (put the cache tag on the response you feed to cache api) and that works, but it doesnt present any better options for cache storage
Jthanks Erisa – that makes sense! I guess I was imagining it as the worker itself (https://github.com/cloudflare/next-on-pages/blob/main/packages/next-on-pages/templates/cache/adaptor.ts) somehow leveraging Cloudflare Cache from fetch e.g.
but there are quite a few incorrect assumptions on there from my own mental model on Cloudflare Cache as you're pointing out AND how actually next-on-pages manages to make it work atm
fetch(somewhereInternal, { cf: { cacheEverything: true, cacheTags: [...] })but there are quite a few incorrect assumptions on there from my own mental model on Cloudflare Cache as you're pointing out AND how actually next-on-pages manages to make it work atm
Ethe blocker from fetch based caching working here is that a worker (or pages function) cant fetch itself in a way where the fetch cache works (it wont on service bindings for example, and those dont work w/ pages anyway) unless you get a weird setup where you host on one domain that then fetches another domain, which is a tad messy and not that easy to support as part of a framework (but i guess theoretically doable)
Jexport const SUSPENSE_CACHE_URL = 'INTERNAL_SUSPENSE_CACHE_HOSTNAME.local';brb signing up for this domain and putting it on Cloudflare as free zone
TAs for next-on-pages, is it possible to edge cache pages for non logged-in users?
AI updated my project and started noticing "Too many subrequests" errors in prod. I made a branch and did some testing, and found that this was caused by updating next from 14.1.0 to 14.2.13 - if I go back to 14.1.0 with the same code then this error doesn't happen. is cloudflare aware of any next change that would be causing this?
TQuestion for the crowd: what is the expect/your median CPU time? (Workers & Pages > project > Function metrics)
I'm curious if mine is high relatively speaking - it seems higher than Vercel but their usage dashboards are a bit difficult to use.
Also, has anyone managed to get
I'm curious if mine is high relatively speaking - it seems higher than Vercel but their usage dashboards are a bit difficult to use.
Also, has anyone managed to get
__NEXT_ON_PAGES__KV_SUSPENSE_CACHE
Aok actually it's weirder than this
Ai noticed that for some queries in my site, it is still getting that error, even on the older next version. (all queries do the same number of subrequests, and it is much lower than 50)
AI made a new branch at the commit I was at before today and deployed it.
Aand this new deploy, with the identical commit as a deploy from a few months ago, gets this "too many subrequests" error on some queries. the older deploy is fine
Aso something seems to have changed in how cloudflare handles these deploys, breaking my site
Ai rolled it back to the old deploy for now, since that's the only way for it to not error
JWe added KV because, well, people asked for it, but my personal opinion is that you should be using cache api since it aligns better with the expected behaviour of the fetch cache (e.g. being regional). And I'm also of the opinion that the eventual consistency of KV completely kills the ability for the cache to work properly because every change to the cache manifest takes 60 seconds to propagate which will mean it'll probably change every request when it shouldnt be
Jare you using the same next-on-pages version in both the old and new deploy?
Aprobably not, it isn't a dependency in the project, and the build command is
pnpm dlx @cloudflare/next-on-pages@1would you be able to try a deploy with next-on-pages version from when your last deploy worked and see if that makes the difference
Aok, it was april 5 so i'll try 1.11.0
Jthat's irritating - there's been a lot of changes since then...
Jcan i be really annoying and ask if it's possible to increment the version until you get to the one that introduces the problem?
Asure
Jthanks
A1.11.3 works fine
1.12.0 has the error
1.12.0 has the error
J
Jdo you use the fetch cache?
Ai don't have any code for controlling caching, but data does get cached. so i assume next does?
Athis is an app router project that has about 20 components that each do a fetch request
Ji'm thinking it's possibly something along the lines of those fetch requests might have been being cached before but one of these two fetch cache changes resulted in them not being cached either as long or at all and therefore leading to more requests happening
Awhether cached or not, there's no reason why one page would do more than ~20 subrequests 
Ji agree, but that's the only thing related to those changes that would make sense
INow that cloudflare has joined OpenNext, do we get a seperate channel for that or do we keep discussions around that in next-on-pages / next channel? https://blog.cloudflare.com/builder-day-2024-announcements/#cloudflare-joins-opennext
The Cloudflare Blog
To celebrate Builder Day 2024, we’re shipping 18 updates inspired by direct feedback from developers building on Cloudflare. This includes new capabilities, like running evals with AI Gateway, beta products like Queues and Vectorize graduating to GA, massive speed improvements to KV and Workers AI, pricing simplifications, and bringing key featu...
TSounds like OpenNext + Workers Static Assets will replace Pages so probably warrants a separate channel?
we plan to bridge the remaining gaps between Workers and Pages and provide ways to migrate your Pages projects to Workers.
Jthe blog post says to use the OpenNext Discord, so you should probably head over there for discussion about
@opennextjs/cloudflare
Dimo we need a migration guide from
@cloudflare/next-on-pages@opennextjs/cloudflare
BIs Cloudflare going to drop support for next-on-pages after OpenNext announcement? If so, a migration guide would be useful
SIs it correctly understood that ISR is not supported by next-on-pages? 
SOh, just saw the above blogpost regarding OpenNext... Interesting!
Jyou can't
KHello, as far as I know axios does not work on edge function right? but why is this api endpoint of mine working just fine
Jmagic
YGitHub
The goal of this PR is to add a minimal fetch implementation that provides the same capabilities as the xhr adapter without major changes, allowing it to be released in Axios 1.x.
To avoid breaking...
To avoid breaking...
CHello, trying to learn and try deploying an existing nextjs project to pages, I'm following the ssr guide. But I got a question.
In the docs, a library
In the docs, a library
@cloudflare/next-on-pagesSnyk Advisor
Find the best open-source package for your project with Snyk Open Source Advisor. Explore over 1 million open source packages.
Tnpm audits typically show a lot of stuff due to how many dependencies npm packages have (you'd need to double check the source of the 11 vulerabilities)... Majority of it (esp for dev deps) is not applicable. Most of the time you just need to re-generate the lock file which updates downstream dependencies
PAre Next.js Workers more expensive than Next.js Pages? Does one HTTP request equal one worker request?
Jit's the same
JPages uses workers
