Ghost 2.21.0 - Very slow after upgrading, 8 Second TTFB

When loading my site after upgrading to Ghost 2 I’m noticing significant load times and it seems to be on initial load & routing that’s the issue. I’m seeing a 8+ second TTFB(Time to first byte) and 20+ second full load. If you view a URL on that same server that doesn’t route through ghost, it loads in 20-80ms so I believe it’s a ghost slowness issue. Is there a way to get to the bottom of what is going on?

  • What’s your URL? http://coreysnyder.me/
  • Server: Amazon T2.micro. Canonical, Ubuntu, 16.04 LTS, amd64 xenial image
  • What version of Ghost are you using? 2.21.0
  • What configuration? Standard
  • What browser? Chrome
  • What errors or information do you see in the console? None
  • What steps could someone else take to reproduce the issue you’re having? Load the site and view the TTYL

Would be great if I could get some technical support to point me in the right direction to figure out what’s wrong. Is this a known issue?

It’s not a known issue. First place to look would probably be your Ghost server logs, they should show how long it look for the Ghost process to generate the response once it received it from your nginx proxy, you should then be able to compare with your nginx server logs.

I have solved this issue. Here’s what was up. I had two blogs running on the same E2 T2.Micro instance. The problem was that the instance was using up all of its CPU credits due to high CPU usage on the box. This was causing my instance to be SUPER slow at running the ghost blog instances. One ghost instance was fully updated, one was old 1.18.4. I didn’t really use the old instance anymore so I stopped it. Immediately CPU usage fell dramatically, to ~1% or less from a steady 11-18%. When I did this, my CPU usage credits started to build back up, and page load times went from 18+ seconds to <200ms TTFB.

Were older ghost servers CPU hogs and this got fixed? Or possibly this old instance I had running was in a bugged out state?

One thing still has me scratching my head. :thinking: I logged into the box and ran ghost stop <name> and then ghost status and I see that the old ghost isntance has been stopped. But when I hard reload the page that the ghost instance was running on, its still delivering the blog. I cleared cache, and disabled network cache, and yet it still seems to be alive somehow. I re-ran ghost status to verify it’s still stopped. CPU Usage is still at 1%. Any idea how/where this is still alive?

Primary Updated Ghost Instance: http://www.coreysnyder.me
Secondary Instance which shouldn’t be running: http://www.coreysnyder.me/buckeyefpv/

@Kevin Do you guys have a recommended Amazon instance size for Ghost to run on?

Not that we know of, it sounds like somehow it got into a bugged state :confused:

Any idea how/where this is still alive?

Is it possible the page is being served from a cache that you have in front of your server? Cloudflare/nginx/etc? Does your theme include a service worker?

Do you guys have a recommended Amazon instance size for Ghost to run on?

The only recommended server specs we have are listed in the self-hosting docs Hosting a Ghost publication - Fully-managed PaaS & self-hosted. 1GB minimum is recommended for install because yarn/npm is memory hungry, in general use Ghost will typically use ~100MB of memory but can peak under heavy load and usage will vary depending on the size of your site.

1 Like

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.