My site is fully accessible globally, but, is now completely inaccessible in most parts of Australia and New Zealand.
What steps have you taken to resolve the issue?
We have attempted to identify the issue, and (while I may be wrong) feel like it’s on Cloudflare’s end? Curl to the origin IP directly with the host passed in the affected regions work fine, but accessing it via the domain? Nope. It has been ongoing for probably >14 hours, and any attempts to identify what’s happening have failed.
What’s most bizarre is how localised and sudden this is, when we have made no changes that would trigger this.
What are the steps to reproduce the issue?
Be within NZ/or on an NZ vpn (Australia should also be able to reproduce, but I believe there is one or two regions within AU the request works fine; so be weary of that).
Attempt to access the site.
The issue has been confirmed through my normal NZ wifi connection, mobile data (of both mine and a colleague’s), as well as an overseas colleague’s VPN in NZ, and also identified the issue as also occurring in Australia. I also have an Australian connection, and was able to reproduce the same issue. Check-host was able to confirm that the site works fine in most places, and another overseas colleague was able to confirm it works in the US.
Would there be any other possible issues? From our end it doesn’t seem to be related to this. At the moment, New Zealand appears to be loudly slowly now with no further action, but outage still exists in Australia.
Yeah, it’s quite inconsistent between 522s and long response times, particularly in Aus (in NZ now I’m mainly just experiencing really long response times). Accessing origin directly works fine, we’ve had quite a bit of a look, unfortunately I feel like this could potentially be a Cloudflare-related issue.
Yeah, I figured. Is there any other possible issues that aren’t related to our end, as, unfortunately, none of what is listed in the 522 tip or Cloudflare’s support articles apply (we have looked at those). We haven’t made any changes recently, and all was working well until the day I originally posted this, so it feels unlikely that this is an error related to our origin.
