Example table
CREATE TABLE rate_limits (
rate_key CHAR(64) PRIMARY KEY,
window_started_at DATETIME NOT NULL,
request_count INT UNSIGNED NOT NULL
) ENGINE=InnoDB;
Application approach
<?php
$identity = $userId
? 'user:' . $userId
: 'ip:' . ($_SERVER['REMOTE_ADDR'] ?? 'unknown');
$key = hash('sha256', 'login:' . $identity);
$limit = 10;
$windowSeconds = 300;
// In production, perform this update atomically in a transaction
// or with a single SQL statement designed for your MySQL version.
$stmt = $pdo->prepare(
'SELECT window_started_at, request_count
FROM rate_limits WHERE rate_key = :key'
);
$stmt->execute(['key' => $key]);
$record = $stmt->fetch();
// Apply an atomic insert/update or transaction here, then reject
// requests whose count exceeds the configured limit.
Important
This is a schema and flow illustration, not a complete atomic limiter. A production limiter must increment counters atomically, define window-reset behavior, and use a trusted client-IP strategy behind proxies. Redis is often a better fit for high-volume rate limiting.