This is an ongoing issue which has been happening for the last couple of months, lasting at least a week each time. It’s possible that the issue is ISP related (EE) but it is impossible to escalate issues to the appropriate technical support team with my ISP, so I thought I would post here in case the issue is something on Cloudflare’s end.
To summarise, packet loss is observed for connections to sites/applications which proxy their traffic through Cloudflare. In addition to packet loss, requests are incredibly slow to respond, e.g. loading a Discord channel may take several minutes and embedded media such as image attachments also take a long time to load. The issue seems to affect all Cloudflare services where DNS Proxy is used including e-commerce sites and personal services I host myself.
MTR shows last hop packet loss when the following route is used: 188.8.131.52 → 184.108.40.206 → Cloudflare IP.
No last hop packet loss is observed when the route is 220.127.116.11 → 18.104.22.168 → Cloudflare IP.
Other EE customers are also experiencing this issue. I have a thread in EE’s Community Forum which includes speed test results and MTR from other users: https://community.ee.co.uk/t5/Broadband-Landline/CloudFlare-Throttling-Issue-FTTH/td-p/1285998.
same issue here, I also have screenshots from cloudflares speedtest website (i cant provide links to it as the forum wont let me) that confirm the issue is only on EE and possibly only the fibre network as i didnt have any issues on ADSL, also worth noting is that i had to run the test multiple times to get it to progress that far most of the time it shows 10mbps or slower speeds and doesnt progress to running any further tests.
I’m having the exact same issue also on EE. Today the connection is failing all together to even reach Cloudflare services at all. I’ve contacted the ISP and they have asked me to try and reach out to Cloudflare for some further information but it seems impossible to locate where to contact Cloudflare directly on this as its affecting a huge range of IP addresses on the EE network.
There appears to be an issue with one of the ways EE connects to our network. This appears to be separate from a previous issue with the same connection last month that was ultimately rectified on their end. The relevent engineers at EE are aware of the issue and are in contact with our engineers.
While they get to the bottom of this, we have disabled the problematic connection from our end. As we have other means of interconnecting at the same location, this should be entirely transparent to the end-user and resolve the degradation of service you have seen. Thank you for reporting this and our sincere apologies.
Thank you Jon. I’m glad both ends have become aware of the issue as this was very difficult for us to get in touch with anyone on either end who would understand the issue, let alone be able to fix it!