I've been doing some work to speed up my site load time, using various tools like GTMetrix, Pingdom, PageSpeed, etc.
The slowest part by a good margin is the initial server response:
I tested some other sites from the Showcase section, and Discuss itself.
Discuss - 350ms
Seek Adventure - @MikeJones - 750ms
DevNL - @JasperVriends - 500ms
Sit. Breathe. Be - @010101 450ms
Average Actuary - me - 900ms
Glass Kraken - me 350 ms
For Average Actuary, I am using a VPS with 2 GB RAM, 1 vCPU, 60 GB SSD. That costs $10 per month with Amazon Lightsail. I have Glass Kraken set up with 4 GB RAM, 2 vCPUs, 80 GB SSD which costs $20 per month and I'm getting about 350ms there. Though, that site has less content, fewer extensions, etc.
Time for some questions!
- Has anyone explored how various server sizes impact the initial server response time? Is there a cliff? That is, no benefit going beyond a certain size?
- Are there things within my control, as far as my unique implementation of Flarum goes, that I can do to speed up the server response time (e.g. avoid certain extensions, etc.)
- Are there things that will speed up the server response time with future releases?
For what it's worth, I also checked the server response times of a few Discourse implementations, and the results seem similair.