Elastic Search Logging Using NGINX Upstream

[Anthony Parsons]

Does anyone know how to get elasticsearch.log to log everything when using NGINX proxying ES as an upstream server?

Setting /etc/elasticsearch/log4j2.properties to 'error = debug' doesn't even work, as all the ES data is being proxied via NGINX... so the NGINX logs just tell the packets received / sent, they don't contain ES info.

I have an issue in my post data that causes ES rebuilding posts to fail, but without ES logs telling me exactly why it failed...

Anyone have experience with this scenario?

 

[Anthony Parsons]

Maybe to more specific, is how to get Elasticsearch data into the log stream?

I have the basic for time and action:

log_format upstreamlog '[$time_local] $remote_addr - $remote_user - $server_name to: $upstream_addr: $request upstream_response_time $upstream_response_time msec $msec request_time $request_time';

Logging to:

access_log /var/log/nginx/access.log upstreamlog;

 

[Xon]

@Anthony Parsons the variable you are looking for is '$request_body' (you probably want client_body_in_single_buffer on too). This this stackoverflow post about it http://stackoverflow.com/questions/...g-the-post-request-body-instead-of-with-nginx

 

[Anthony Parsons]

Thanks @Xon, will give that a go and see what happens.