I’m not sure if discourse had cache or not, i’ve tried using loader.io to test my forum and it as fast as using cache but i can’t find cache software through wappalyzer, so what cache does discourse use? does it built in cache? and does it possible to add third party cache?
What do you mean by “cache”? Discourse benefits from use of a CDN, for the standard reasons – shared web / JS assets that don’t change much can be cached either on the user’s device, or by a local CDN node geographically closer to the user.
What i mean by cache is just like ‘Varnish cache’ that installed on a website. does discourae has built it cache like that?
Oh so all our asset on forum are being hosted on cdn not on our local host?
No, because Discourse is for the most part serving dynamic content. Varnish as in in-memory cache of static content makes no sense here.
That depends, if you’re self-hosting have you configured one?
are you sure? stackoverflow using varnish, and almost every minute there a question
Yep, very. There’s nothing for Varnish to accelerate.
Stop thinking about Discourse as a website and start viewing it as an application. The use cases are very different.
View your F12 network tab in the browser if you don’t believe me. Watch the requests that come through.
ah so discourse its very lightweight, that doesnt even need cache, right?
im just curious that im using plain wordpress on same vps and “brute force” my site, the success rate is about 8% but when using discourse its all request are received just like using cache on my wordpress
WordPress is designed to serve a high volume of pages with little to no change to a high volume of users. Write little, read lots which requires little resources to do.
You can use plugins to build gzipped static copies of pages, database queries can be cached in-memory and the whole site can be front-ended with a product such as Varnish.
Most Wordpress sites serve the majority of their sessions to logged-out users.
Discourse on the other hand serves views tailored to the individual. The majority of the users are logged in. Write frequently, read frequently which comparatively requires lots of resources.
There’s really no comparison.
Your test is based on a flawed premise, particularly when you consider this:
Do you load linked resources/assets?
Their test doesn’t even load Discourse correctly.
Also, yes, Discourse does have a cache of full page responses - for logged-out (anonymous) users, which it sounds like this “load test” is doing.
The Discourse team has also benchmarks that directly exercise the known-expensive code in the critical serving paths, you can find them if you search around on here.
Please sum it up, does discourse have cache on their engine? that’s the only answers i need.
… We use Redis as a cache and for transient data. …
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.