Logging level not respected by Haproxy

[mateuszdrab]

Hi

I am experiencing an issue where the logs of debug level flood Haproxy logs even when the controller is started with --log=info flag. This would not be a massive deal; however, since the logs of haproxy are stored for longer retention, lots of unnecessary space is used as there are currently over 30k+ decisions and each IP is listed during reload, along with each IP source from an incoming request.

What confuses me, is that according to the documentation - the methods under core are provided by Haproxy itself. I'm currently running haproxytech/kubernetes-ingress:1.10.7, which is the latest.

Is there something I'm missing like a Lua specific config directive or is there a bug with Haproxy that I need to report to them?

Thanks