Struct LockoutPolicy 

pub struct LockoutPolicy { /* private fields */ }

Per-principal lockout policy. A “principal” is typically the user id (so a stuffing attack against one account can’t lock out others) or, before the user is known, the client IP.

Implementations

impl LockoutPolicy

pub fn new( backend: Arc<dyn RateLimitBackend>, max_attempts: u32, window_secs: u32, ) -> Self

max_attempts failed logins allowed per window_secs sliding window.

pub fn fail_open(self) -> Self

If the backend is unavailable, allow the attempt instead of blocking (default is fail-closed — deny). Use only when availability outranks the brute-force risk for this deployment.

pub async fn check(&self, principal: &str) -> Result<()>

Count this login attempt against the principal’s sliding window and reject once the threshold is crossed.

The backend exposes only an incrementing hit (no peek), so — exactly like the framework’s DistributedRateLimit fronting /auth/login — this counts every attempt, not only failures. Call it once at the top of login, keyed by a principal that identifies the target before the user is known (e.g. tenant + email, or the client IP), so unknown-user probing is throttled too and the response can’t enumerate accounts.

Returns IdentityError::AccountLocked when the window is exhausted, or IdentityError::Unavailable when the backend is down and the policy is fail-closed (the default).