Fix: Make sure that missing supervisors are correctly displayed when using multiple environments

[PrinsFrank]

When displaying active/inactive supervisors for a master supervisor that is started with a different environment, the provisioning in the overview is done with the configuration of the environment the horizon endpoint runs on, causing a potential mismatch when using different environments. This PR passes the environment to the master and uses that variable to display a correct overview. (See discussion in PR #1286 for a reproduction)

[taylorotwell]

Why would the environments be different?

[PrinsFrank]

Why would the environments be different?

In the linked PR @bjarn had an example where there were multiple environments with different supervisor configurations.

I've seen this setup at more larger companies, and it provides another dimension of scalability. For example, you can use 3 default configured pods running the default production supervisors, but then when there is a workload on a specific queue that is suffocating the other queues, you can add a dedicated pod for a different environment running only that queue.

[
    'production' => [ // Started with "php artisan horizon"
        'supervisor-1' => [
            'queues' => ['default', 'low-prio', 'high-prio', 'index']
        ],
    ],
    'production-index' => [ // Started with "php artisan horizon --environment=production-index"
        'supervisor-1' => [
            'queues' => ['index']
        ],
    ],
]