AWS / CloudFront
- Thread starter zagman76
- Start date
Hi,
A quick search of your posts indicate that the issue may lie with the server configuration eg http and php processing, as well as caches (are the requisite cache add-ons installed and enabled?).
TTFB of more than 10 seconds is obviously really bad. My configuration on AWS running nginx/php-fpm has a TTFB of 0.476s on first view.
A quick search of your posts indicate that the issue may lie with the server configuration eg http and php processing, as well as caches (are the requisite cache add-ons installed and enabled?).
TTFB of more than 10 seconds is obviously really bad. My configuration on AWS running nginx/php-fpm has a TTFB of 0.476s on first view.
My TTFB’s have been between 2-15 seconds (and a few in the 30-45 second range). I finally got caching working (Redis), and that’s helped a bunch.
What I see happening now, is a click (node/forum/thread/etc.), then a really, really long delay, then the page explodes to life. Developer panel shows all things coming in at less than .5s, except TTFB.
What I am baffled about, is that it only started happening with the XF2 upgrade last week. Before then, it was mostly smooth sailing.
I’ve read a bunch of guides on how to fine-tune php/php-fpm/nginx, and it would seem that they are tuned nicely.
When I look at the ELB metrics, it shows back-end latency in the 2-15 second range, which is consistent with what I’m seeing when loading the page in the browser.
On one hand throwing more hardware (i.e.: CloudFront) at the problem is more of a temporary fix, until I can figure out why the problem exists in the first place.
Are you by any chance using Xon's Redis Cache add-on for Xenforo 2? I don't like to overload my installation with add-ons that aren't for big things so I never installed it. And if so, how is it working out for you? My site is small at the moment but it would be nice to decrease page load time
No - I launched a redis cluster in AWS, and attached directly to it via the src/config.php file.
