A user, meglio, is experiencing slow loading on their Discourse instance, running on a $10 VPS at DigitalOcean with 400 users and 50 active users. To diagnose the issue, sam suggests running . /launcher cleanup. codinghorror recommends checking the Digital Ocean droplet graphs for unusual disk or CPU activity, and increasing the droplet size if necessary. Additionally, ljpp advises connecting to the VPS via SSH and running top to identify the issue, suggesting reporting the findings.
Still running 10$ VPS at DigitalOcean, I started to experience slow loading a few days ago. Not only do I understand that the VPS is the minimal recommended one, but I can also see there is lack of space. Running just around 400 users and just 50 of them being active, I’m confused whether it is disk space and/or VPS processing power that is a bottleneck.
What is the best way to check and know for sure?
Try running:
./launcher cleanup
The general answer to this question is to look at your Digital Ocean droplet graphs:
(above is the last year of http://discourse.codinghorror.com)
Anything weird with Disk or CPU, super high peaks, you need to upsize your droplet.
Remember when you upsize the droplet to increase the number of unicorn workers, and the amount of memory allocated to Postgres.
@meglio connect to your VPS via SSH, run top and see whats happening. Maybe even report here.
