Built-in checks

The package ships fluent checks for the common cases, each with a default notification so you do not have to write one:

use RoundlyConsulting\Alerts\Checks\{CacheCheck, DatabaseCheck, HttpPingCheck, QueueCheck, StorageCheck};
use RoundlyConsulting\Alerts\Facades\Health;

Health::check(DatabaseCheck::make('mysql'));
Health::check(CacheCheck::make('redis'));
Health::check(QueueCheck::make()->onQueue('emails')->maxSize(1000));
Health::check(StorageCheck::make('uploads')->minimumBytes(2 * 1024 ** 3));
Health::check(HttpPingCheck::make('https://api.example.com')->timeout(5)->expectStatus(200)->slowerThan(800));

Options and defaults

use RoundlyConsulting\Alerts\Checks\QueueCheck;
use RoundlyConsulting\Alerts\Checks\StorageCheck;

Health::check(
    StorageCheck::make('local')
        ->minimumBytes(5 * 1024 ** 3)   // fail below 5 GB free
        ->warnBytes(20 * 1024 ** 3)     // warn below 20 GB free
);

Health::check(QueueCheck::make('redis')->onQueue('default')->maxSize(500));

• StorageCheck reads free space at the disk’s local root path, so point it at a local disk. Without warnBytes() it warns below twice the minimum.
• HttpPingCheck’s timeout() is the HTTP request timeout. It is separate from the per-monitor timeout() on the scheduling builder.
• Each result carries useful meta: connection, store, size and max_size, free_bytes and minimum_bytes, or url, status and latency_ms — plus error when an exception was caught.
• Result messages come from the alerts::checks translation file, so you can reword them after publishing the translations.
• Each built-in registers under the key in the table. To run several instances of one built-in — two databases, three endpoints — give each its own key with as() (see Registering checks).

Scheduling a built-in check

Register the configured instance once, then schedule it by class for each owner. The row stores only the key; at run time the registered instance, with its URL, connection or thresholds, does the work:

// Register the configured instance once, in a service provider…
Health::check(HttpPingCheck::make('https://api.example.com/health')->slowerThan(800));

// …then attach it to every owner that should hear about it.
Health::for($team)->monitor(HttpPingCheck::class)
    ->everyMinute()
    ->failAfter(2)
    ->save();

An instance registered with as() is scheduled by its key instead — Health::for($team)->monitor('mysql_database'). Every built-in check notifies with the bundled HealthCheckFailedNotification, which includes the failure reason, and honours notifyVia() channel routing (see Notifications & throttling).