Possible solution to concurrent instances working?

[Wintersunner]

I've noticed an issue with having multiple jobs queued in environments with more than one instance. Specifically, if a job cycle exceeds 70 seconds (the default _interval + 10), the lock expires, allowing another instance to start enqueuing jobs, which results in duplicate jobs in the queues.

Using a single instance to avoid this problem is not ideal due to its slow performance and being a single point of failure.

Is there a solution to make rqscheduler horizontally scalable? Not just for failover but to operate simultaneously and increase throughput?