Out of the blue I woke up this morning to all of my sites being down. I thought it was the actual servers (3 dedicated servers that are in a Virginia datacenter), but I’ve been on and off of them all day without any connectivity issues… With addresses that aren’t proxied seem to work fine, but the ones that are proxied through Cloudflare are hit or miss… Sometimes they come up, sometimes they throw a 520 error, sometimes a 522 error, sometimes they pull up the Cloudflare “Always On”…
Thanks for the “pause Cloudflare” suggestion! I just did that for one of the sites and it was lightning quick (like usual)… The issue is that I can’t leave it like that for security reasons
There was something at the hosting center earlier (swapping something out at one of the routers), but it was supposed to have a backup and no loss of connectivity… That was fixed this morning, but the sluggishness persists… I guess maybe it is on the Cloudflare end if it went away when I paused Cloudflare…
If you sites are slow and get always online pages and combine that with 520/522 errors, then 90% sure it’s because your origin server is overwhelmed, crashing and/or unable to handle the traffic or load from visitor requests. You’d want to systematically rule out each possibility starting with
database queries/slow database MySQL queries
slow or unoptimised web server/PHP server requests
web application configuration i.e. Wordpress plugins forum addons or or web app related slowness
your origin server configuration issues i.e. origin firewall or HTTPS SSL certificate configuration
web host specific issues i.e. web host firewall or security/DDOS mitigation systems in place
It could be any combination of the above listed issues causing 520/522 errors. I recently fixed one of my clients 520/522 issues which they had up to 200,000 HTTP 520/522 status errors per day which accounted for ~1.7% of their total traffic and they had a combination of all the above listed issues that needed fixing. Alot of it was down to poor optimization and configuration by their previous hired system admin help!
Yes, I’ve been through all of those. They’re the first things I start digging into when things come up to make sure it’s not something that I’ve done It’s a very mature custom built CMS (C# & MS SQL) that’s been in production for over 14 years, so the vast majority of those issues have been fixed over the years (tweaking performance, rewriting old code, etc.). All of the sites have also been on Cloudflare for several years without issue and not much has been changed in those settings. Traffic levels are normal, etc… It just started cropping up about 9am this morning out of the blue.
I do have some timeouts here at my office, which tells me there are some issues in the backbone here in Utah, but they’re fairly minor and also I’m testing from other locations in other states.
Anyway, thanks for the suggestions! I’ll keep hunting