My client wants to track down the brute force attempts.
To achieve that, he counts the brute force protector warnings out of the logs. I don't know any other way to get such information.
If I set the threshold to 3, I should have 2 free go (not 1) and then the 3rd failed attempt should be logged as brute force attack.
Exactly what I'm saying : the brute force protector should not log all failed logins when we are below the set treshold.
I cannot simply ignore the brute force protector messages as you suggest, because I want to know when I'm facing a brute force attack and act accordingly (phone the customer that his account has been hacked or that suspisious activity take place with his account)
Tracking down user_temporarily_disabled statuses may be a way, but I need to build up some watchdog mechanism, the locked-down customer being active again after some configurable time.