A client has a lot of traffic, and I’ve tried messing with offloading assets to S3, trying different memcache/redis solutions, advanced post cache etc, but I’m wondering if someone has a concrete setup they can recommend, along with how to efficienctly load test it all.
Both of these are a little dated but:
Also worth listening to: https://roots.io/podcast/ep12/
Amazing resources! Thanks for them, surprised I missed them both.
I can’t specifically comment on scaling Trellis in production, but I have done a bit of this related to a WooCommerce stack recently.
- Make sure you’re using Redis, one of the recommended hosts for Roots stuff, Kinsta, charges $100/mo for it, but some hosts have it included (such as Pantheon). I’m using the plugin on the .org repo by Pantheon with Kinsta currently.
- Is your caching solution working? What percentage of cache hits/misses are you getting?
- New Relic can help you diagnose where issues might be and some queries that might not be performing.
- Are you putting things in transients?
For load testing there are tons of options, but I’m working on something right now and hope to write a post on the blog about it when I’m complete.
This is a very broad topic, but only some things are going to apply to you based on the hosting solution you’re currently using. Also the type of traffic you’re getting. In my case, I also just needed much more power.
Thanks for all of your points, I do need to somehow check my wp-redis cache hits/misses. Would love to know when you’re done with the load testing option. loader.io is what I use currently.
I actually meant page cache hits/misses. But redis hits/misses can be easily checked with Query Monitor.
I think we’re planning on using Load Impact for load testing.
How can you check that hit miss rate?
Also could you share the specific configuration you’re running please