Comment by dekkers

What the article writes about scalability also applies to availability. Does the queue need 99.999% or 99.9999% uptime? Or is the Service Level Objective actually 99.99%, 99.9% or even 99.5%?

With 99.99% you can have 4 minutes of downtime a month. If failover to a hot standby takes a minute then that shouldn't be a problem to achieve a 99.99% uptime SLO.