How brute force protection works
A brute force attack is unsophisticated by design. Software submits your login form again and again with different passwords until one works. A related and more effective version, credential stuffing, skips the guessing and tries username and password pairs already leaked from other services, on the assumption that people reuse them.
Protection makes repetition expensive. Rate limiting slows a source that submits too quickly; lockouts block an address after a small number of failures; a challenge such as a CAPTCHA forces a human step; and two-factor authentication means a correct password on its own is no longer enough. Most tools combine several, and the important design point is that they target the attempt, not the attacker, since addresses are cheap to change.
Why it matters
The obvious risk is losing the site. An administrator account is complete control: content, redirects, injected scripts, everything. Sites are rarely singled out for this — automated scanners work through lists of addresses, and any WordPress installation is a target simply for existing.
The less obvious cost is performance. Every login attempt runs code and queries the database, so a sustained attack behaves like a small traffic flood. On shared hosting, which is what most small businesses in Nepal use, that is enough to slow the site for real visitors or trip a resource limit, and the owner sees an unexplained outage rather than an attack. The same applies to older interfaces such as XML-RPC, which can accept many credential attempts within a single request and is often left enabled on sites that never use it.
Where brute force protection goes wrong
Lockouts based on the username rather than the source can be turned against you: anyone can lock out your administrator account deliberately by failing to log in as them. Lockouts based on the source alone are also imperfect, because a distributed attack simply rotates addresses and because several of your own staff may share one office connection.
Hiding the login page is widely recommended and widely misunderstood. It removes noise from automated scanners, which is genuinely useful, but it is not security — anyone who finds the new address is exactly where they started. Relying on it while keeping a weak password is the wrong trade. The remaining mistakes are practical: adding a challenge that also frustrates real users on enquiry forms, forgetting to protect the API and XML-RPC alongside the login form, and locking yourself out with no recovery route through your host.
How to act on it
Do the cheap, decisive things first. Give every account a long unique password from a password manager, remove accounts nobody uses, and give people the lowest role that lets them do their job — an editor does not need administrator rights. Then add two-factor authentication on every administrator account, which defeats guessing and stuffing outright.
After that, limit attempts with a sensible lockout, protect the API and XML-RPC or disable XML-RPC if nothing you use depends on it, and keep a way back in through your hosting control panel in case you lock yourself out. Watch the failed-login log occasionally; a sudden rise usually explains a slow site better than any speed report. Where the site earns money, fold these checks into a standing website maintenance routine rather than revisiting them after an incident.
