Cockpit capacity

#1

I really like Cockpit.
That’s why I switched some of my projects to Cockpit. But I haven’t tried a big web application with Cockpit yet.

Currently I have a task to build a hotel management app for my client. And I guess 1k+ users will be using this app everyday.

How do you think on this?
or any other suggestion?

Thanks in advance.

#2

It depends on what you plan to expose to the end users, Cockpit is an Headless CMS, so I’m assuming that you’ll have an application consuming the API, if so, you should be fine, some aspects to consider:

  • Use MongoDB
  • Use cache in the BE (e.g. using the simple response cache or Redis add-ons)
  • Minimize API requests in the FE
1 Like
#3

Thank you so much @pauloamgomes , I’ll try it with Cockpit.

#4

You can also test your app with apache stress test. Cockpit is doing very well actually. I got 700 api response/sec on my virtual machine (2gb ram, 4 core cpu CentOS) without API cache add-on.

Example usage:
ab -n 5000 -c 200 http://yourdomain.com (without SSL)
abs -n 5000 -c 200 https://yourdomain.com (with SSL)

1 Like
#5

Thanks @KafaKonforu, can you provide more details how are you testing? like what api requests you are doing (e.g. fetching all entries of collection x), and complexity of the collections (number of fields, number of linked fields, etc…). Would good if we can have some kind of reference for benchmarking.

Think it should perform better than most of identical CMSs mostly because the codebase is small and it has few dependencies.

#6

Firstly i am using cockpit for mobile app back-end. I was fetching an entry from collection which has 6 fields and about 600 entries. It has almost identical performance on simple query like user auth etc.

Compared to others;
CodeIgniter has 280-300 request/sec on simple api query with cache,
Laravel has 320-350 req/sec with cache
Wordpress is around 120-140 req/sec with cache, its impossible to complete stress test without cache :slight_smile:

Surprisingly, simple response cache add-on slows down cockpit performance on ssd machine. The score was 700req/sec without cache, 630 req/sec with cache. Interesting…

1 Like
#7

And using redis-cache addon?