Cloudflare is unstable from few days and its a disaster with 520 error from cloudflare

crazy that this is not solved after so many months now. Stil getting these errors across 20 different sites on a dozen different servers… the past days more than ever.

I only have a single site and not insane traffic so maybe not much data to go from, but I’ve never ever seen this error and never heard about it from any user. What’s the difference? Is it because we’re on a paid plan?

Also seeing 0x 520 status code from 470k requests in the past week in the analytics tab.

if moving to paid plan solves the problem let me know folks
cos free plan is doomed !!

In my extensive testing, I have noticed the 520 errors happen on servers running Apache. It doesn’t matter what version, old or new. Again, this problem never happened before the week of July 22… still happening today. I thought maybe it had something to do with the newest version of Apache, and tried older versions, but nothing changed the random CF 520 errors. And our servers running Nginx appear to not have any errors, that I can find.

The only other distinction I can make is before CF attempted to fix this issue, the errors were more random, to include both randomly clicking a link (page load) and when doing a website action (submit data). Since CF has fixed the issue, the page load 520 errors appear to happen less frequently and the submit data pages continue to have intermittent 520 errors.

I’ve disabled Always Online and disabled Brotli, but nothing solves the problem. And users are frustrated.

Our site with no errors also runs behind nginx.

what’s your apache version? output from command

httpd -V

maybe clues in https://downloads.apache.org/httpd/CHANGES_2.4

I am certain there is no problem with Apache on the server side. The problem is 100% on the Cloudflare side of things. I am uncertain if it has to do with Cloudflare’s interaction with Apache… Cloudflare’s response has been there was a problem with Always Online, but they don’t give any other direction. Frustrating because when you turn off Always Online, the 520 errors still continue, albeit less.

We are running 20+ servers. 20+ are Apache, and 1 is Nginx. The sole Nginx has been the only one that hasn’t seen any problems. And as I mentioned before, we have made sure we have updated Apache, but that did not solve the problem, nor did downgrading to previous Apache versions solve the problem. The reality is on over 20+ servers, we were experiencing absolutely zero errors before the week of July 22 and no updates were made on our end. Cloudlfare made a change around that time, affecting every server.

its almost a month and the issues are worse today
always offline show everytime and the 520 errors haven’t stopped :frowning:
can someone escalate this to cloudflare

Server version: Apache/2.4.48 (cPanel)
Server built: Aug 19 2021 14:52:05

matteo any update from cloudflare ? please push them to respond

its mid OCT and till date no update on this 520 erorr

Ray ID: 69af9ec3aff46ee9

my ticket ID is - 2248400

2 Likes

I fixed it turning keep-alive off on my server (HAProxy). The Cloudflare team said last Sep (Random 520 errors - #24 by homebarista) that the issue with keep alive was fixed, but now I’m starting seeing the issues again. I tuned keepalive off, and the issues gone away.

To everyone, please turn keep-alive off and leave a comment in this thread if the 520s errors gone away, to let know to the Cloudflare team if the issue with the 520s it’s related with the keep-alive option.

Thanks,