We’ve been trying to investigate some 522 issues, we know some customers are seeing it, but we’re also seeing it each day in the office. It’s intermittent and usually just a single request and then ok for hours, until it happens again.
Cloudflare support have provided some details from their logs based on the approximate time and ray id I sent in a screenshot, but neither Cloudflare or our web host are accepting responsibility for the issue and blaming the other. Our local server logs don’t show any errors around the time of these requests, we’ve already confirmed the list of Cloudflare IPs are allowlisted on any firewall/application filtering.
Cloudflare support said:
We have a few errors here. Here’s the 3 type of errors we have logged on out system on that rayID time frame:
ConnectionClosed
“timestamp”: “2023-05-19T14:23:09Z”,
“message”: " context: Prematurely before response body is complete",
“host”: xx-hidden-xx,
“request”: “GET /shop/page/3/?orderby=popularity”,
“upstream”: “xx.xx.xx.xx:443”,
“status”: “0”,
“error”: {
“source”: “Downstream”,
“type”: “ConnectionClosed”,`ConnectionClosed
“timestamp”: “2023-05-19T14:23:04Z”,
“message”: " context: Prematurely before response header is sent",
“host”: “xx-hidden-xx”,
“request”: “GET /shop/weight-loss/”,
“upstream”: “xx.xx.xx.xx:443”,
“status”: “0”,
“error”: {
“source”: “Downstream”,
“type”: “ConnectionClosed”,ConnectTimedout
“timestamp”: “2023-05-19T14:23:04Z”,
“message”: " context: Fail to connect to xx.xx.xx.xx:443 cause: Connection timed out (os error 110)",
“host”: “xx-hidden-xx”,
“request”: “GET /wp-content/plugins/paypal-for-woocommerce/assets/js/angelleye-admin-v2.js?ver=1684506173”,
“upstream”: “xx.xx.xx.xx:443”,
“status”: “522”,
“error”: {
“source”: “Upstream”,
“type”: “ConnectTimedout”,
“context”: “Fail to connect to addr: xx.xx.xx.xx:443”