Throttle middleware affects the entire application scan about laravel-seo-scanner

[Baspa]

Hmm I don't think that's too low at all. But I think because it's grouped we reach the limit too fast. I'm currently thinking about building some throttle mechanism into the package so you can configure the maximum amount of requests done per minute. You're probably not the only one using the throttle middleware so we should find a solution for this.

I will try to find a good solution for this, if you have any ideas in the meantime please let me know 👍🏼

[Baspa]

Only one request is made for each url, so it should only be visited once by the scan. Can you tell me how many requests can be made each minute to your routes and how many requests you receive within that minute?

[MuriloChianfa]

apparently the throttle middleware groups some routes when used with the Route::group function, thus forcing the route group.
I defined the access limit as 30:1 (30 requests every 1 minute) but I have about 40 routes protected by the same throttle middleware to avoid flooding, do you think that is too low?

[Baspa]

I think the problem is with the isBrokenLink method which makes a request with curl to determine if the link is broken:

if (! function_exists('isBrokenLink')) {
    function isBrokenLink(string $url): bool
    {
        $statusCode = (string) getRemoteStatus($url);

        if (str_starts_with($statusCode, '4') || str_starts_with($statusCode, '5') || $statusCode === '0') {
            return true;
        }

        return false;
    }
}

if (! function_exists('getRemoteStatus')) {
    function getRemoteStatus(string $url): int
    {
        return cache()->driver(config('seo.cache.driver'))->tags('seo')->rememberForever($url, function () use ($url) {
            $ch = curl_init($url);

            $options = [
                CURLOPT_RETURNTRANSFER => true,
                CURLOPT_HEADER => true,
                CURLOPT_NOBODY => true,
                CURLOPT_TIMEOUT => 10,
                CURLOPT_FOLLOWLOCATION,
            ];

            if (app()->runningUnitTests()) {
                $options[CURLOPT_SSL_VERIFYHOST] = false;
                $options[CURLOPT_SSL_VERIFYPEER] = false;
                $options[CURLOPT_SSL_VERIFYSTATUS] = false;
            }

            curl_setopt_array($ch, $options);
            curl_exec($ch);

            $statusCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);

            curl_close($ch);

            return $statusCode;
        });
    }
}

Possible solutions

• We could use Guzzle Advanced Throttle. I don't know if using Guzzle slows down making the requests, I should check that first.
• Use some sort of queue mechanism in the package which can limit the number of requests made at any given time.