522 Connection timed out

[nathaniel]

All of a sudden today I cannot access my Cloudflare hosted web apps (I get 522):

https://beta-api.fitchek.com/v1/corporate/status
https://beta-wellness.fitchek.com/

I have made no configuration changes. I can access this beta server via SSH command line, and I have rebooted the server, but no effect. I was able to access everything (also see below) normally yesterday, and to my knowledge nothing has changed configuration wise locally, on beta, or production.

Another (related?) strange problem I am having is that my production web apps, being hosted on AWS, with the DNS set by Cloudlfare. I CANNOT SSH to the EC2 instance on AWS from the command line, I get “Connection timed out”

ssh: connect to host ec2-35-182-173-184.ca-central-1.compute.amazonaws.com port 22: Connection timed out

but the web apps ARE still running:

https://api.fitchek.com/v1/corporate/status
https://wellness.fitchek.com/

Can anyone help. I have checked with my hosting provider Rogers Canada, and they have confirmed that there are no issues from their end, so the problem must be somehow related to Cloudflare, unless I am missing something.

[Cyb3r-Jak3]

My guess is something happened to your AWS. 522 means that Cloudflare is unable to reach your server and the lack of ssh leads me to believe that something happened on the host.

[nathaniel]

Except that

a) my beta server is not hosted by AWS, it is hosted on a local testing server with Rogers and Cloudflare, and those are the websites that are causing the 522. I am able to SSH to the beta server without isusue. All DNS records for beta urls are A records in Cloudflare.

https://beta-api.fitchek.com/v1/corporate/status
https://beta-wellness.fitchek.com/

b) My AWS web sites, have the exact opposite issue, the are urls are loading properly, but I cannot SSH into the EC2 instance.

ssh: connect to host ec2-35-182-173-184.ca-central-1.compute.amazonaws.com port 22: Connection timed out

https://api.fitchek.com/v1/corporate/status
https://wellness.fitchek.com/

All records for AWS production urls are CNAME records that alias to the AWS EC2

ec2-35-182-173-184.ca-central-1.compute.amazonaws.com

[Cyb3r-Jak3]

I would check the logs on your testing server to see what errors are occuring.

Are the CNAME orange or grey clouds?

[nathaniel]

CNAME are orange clouds. Nginx logs error and access show nothing.

[Cyb3r-Jak3]

Has the IP changed? It is the only thing that comes to mind which wouldn’t show the logs.

Also, are you using the CNAME record to ssh or directly to the AWS instance.

[nathaniel]

You nailed it. We must have had a power outage overnight that caused us to lose the lease on the IP and we were assigned a new one by Rogers. Thanks for the help!