Issue Amazon AWS Route53 Health Check Server Level Access Requests are Blocked with a '400' error code

[Ehud]

Server operating system version

nginx version: nginx/1.24.0 Server version: Apache/2.4.57 (Ubuntu) Server built: 2023-04-08T12:56:02

Plesk version and microupdate number

Product version: Plesk Obsidian 18.0.53.2 OS version: Ubuntu 22.04 x86_64 Build date: 2023/06/19 12:00

Hi,

As part of configuring DNSSEC on the AWS side, for a Plesk run server, the AWS guidance calls to setting an Amazon-Route53-Health-Check-Service CloudWatch Alarm.

Apparently, this Cloud Watch, access the instance on server level, and in our case, is rejected with a '400' error code. This rejection occurs by an nginx proxy server standing in-front of an Apache2 server,

Below is a log example:

2023-07-03 12:26:38                54.183.255.151 - - "GET / HTTP/1.1" 400 248 "-" "Amazon-Route53-Health-Check-Service (ref ref_number_code; report http://amzn.to/12345)"                server_nginx_access_log
2023-07-03 12:26:39                54.255.254.247 - - "GET / HTTP/1.1" 400 248 "-" "Amazon-Route53-Health-Check-Service (ref  ref_number_code; report http://amzn.to/12345)"                server_nginx_access_log
2023-07-03 12:26:39                54.243.31.247 - - "GET / HTTP/1.1" 400 248 "-" "Amazon-Route53-Health-Check-Service (ref  ref_number_code; report http://amzn.to/12345)"                server_nginx_access_log
2023-07-03 12:26:39                176.34.159.247 - - "GET / HTTP/1.1" 400 248 "-" "Amazon-Route53-Health-Check-Service (ref  ref_number_code; report http://amzn.to/12345)"                server_nginx_access_log
2023-07-03 12:26:39                54.252.254.215 - - "GET / HTTP/1.1" 400 248 "-" "Amazon-Route53-Health-Check-Service (ref  ref_number_code; report http://amzn.to/12345)"                server_nginx_access_log
2023-07-03 12:26:40                54.241.32.119 - - "GET / HTTP/1.1" 400 248 "-" "Amazon-Route53-Health-Check-Service (ref  ref_number_code; report http://amzn.to/12345)"                server_nginx_access_log
2023-07-03 12:26:40                107.23.255.23 - - "GET / HTTP/1.1" 400 248 "-" "Amazon-Route53-Health-Check-Service (ref ref_number_code; report http://amzn.to/12345)"                         server_nginx_access_log

I could not find any reason for a '400' error code. If it was a '403' error code, I could have get an idea where it could come from, but not this '400' error code.


May I ask, if you have any thoughts, how to identify where is the '400' error code is coming from, and how to allow the checking to take place, not being blocked?

 

[Ehud]

BTW - The issue is NOT reflect in the server error log, but in the access log...

nginx
Logs
Global logs:
Access log: /var/log/nginx/access.log
Error log: /var/log/nginx/error.log

 

[Ehud]

It seems I making progress in finding where the '400' error could come from. Apparently Plesk manages addiotnal SERVER level APACHE2 log, which is run not as SERVER level but as DOMAIN level, although it using IP ACCESS and not domain name access...

/var/log/apache2/other_vhosts_access.log

Which, I would consider as some sort of a Plesk BUG for its own, as IP Access IMHO should not be considered domain level.

I can see the logs there start with the line:

example.com:80

which is not reflect on the PLESK GUI log, where I see the report as coming from server_nginx_access_log.

 

[Ehud]

Ok.. I now suspect a regular domain level blocking of numeric referrer was what blocked the AWS health check.

It's strange in my eyes numeric access request even gets to the domain level, and is not considered server level request.