hello,
Nice work! It’s an excellent idea.
but it could be obvisous to be able to tweak not only things in config.php but also in php.ini. For example: opcache and opcache_jit configuration
Thanks a lot for your work
In the past, I’ve ever had to deal with opcache.interned_strings_buffer value in php.ini (especially when there are many apps installed in nextcloud)
And I’ve noticed that the nextcloud instance is more reactive with theses opcache.jit directives in php.ini
I had installed nextcloud on fresh yunohost vps (1cpu, 2G/20G). With only two users. I experienced a lot of internal server errors, increasing the memory footprint fixed it (set it to 200, I don’t know if it’s reasonable since the high setting is >40). But it was extremely slow. I tested it on webpagetest.org which said in the report that it needed 65 seconds to completely load the login page . I removed some apps in nextcloud (I don’t remember exactly, I left active: photos, talk, activities, office, share). And rerun the test and it took only 16 seconds. I don’t know if there more performance tweaks that fit to the hardware config of my server.
I didn’t know about that site. Is there a comparable self-hosted alternative, to your knowledge?
RIPE has Atlas, which offers a global network for ‘is my site up’ tests, and offers small hardware probes to become part of the network.
If there’s some self-hosted package that allows our co-yunohosters to run tests, we could have this kind of distributed monitoring.
Anyway, sorry for totally off-railing the thread
@ljf : looks nice, thank you! I think I once set config.php to read-only (or set a read-only option inside it), as a workaround for another option. If you have no idea what I’m talking about, I’ll try to look it up.
And this one caught my attention, it’s open source and can be self hosted (and it’s not hard)
I tried it on my nextcloud server and it threw me a 0 for compression (and talked about brotli compression, which is compatible with nginx, I’ll check if it’s enabled and try to install it in case it’s not). Other bad results are about bad css and css complexity, things I can’t fix by myself.
Good point @wbk
You can check the remaining apps on the list in case