How to perform queuing and caching using redis for a multi-tenant multi-database architecture application in Laravel 5?

Question

I am working on a multi-tenant multi-database architecture application using Laravel which basically means that every tenant in the application has their own database, own sets of users, resources and so on.

Now I am trying to implement queues and caching in the application, then I am trying to use redis for that. A sample code looks like this:

$mailer->send('emails.welcome', ['user' => $user], function ($message) use ($user) {
    $message->from("[email protected]", "Admin");
    $message->to($user->email, $user->first_name)->subject("Welcome!");
});

This is to send a welcome email to user sign-up. But the queuing is storing all the queues in the same database in the same Redis instance, different tenant emails will get mixed up as far as I can tell.

How do I hook into Laravel 5 and change the queue behavior to either store the jobs for each tenant in a separate database or store extra meta-information about the tenant a particular job belongs to? And then also how would I tell Laravel to parse that extra meta-information and connect to the right tenant database before proceeding with the job?

Answer 1

For appropriate work of Queue system, you need to use your own implementation of \Illuminate\Queue\SerializesModels trait inside your Jobs. Wheech will save and init correct DB connection on Job::__sleep() and Job::__wakeup(). Take a look at TenantAwareJob trait of hyn/multi-tenant package.

For appropriate work of Cache system, you need to use a prefix which depends on the current host. Take a look how hyn/multi-tenant developers recommend to implement this.