How do clients decide which DNS server to use from the list of DNS preferences?

Hi all,

So to summarize my current network setup, I have a Unifi USG gateway/router, and I currently have two pi-hole servers running as DNS servers (yes two). Long story short, I found my old raspberry pi and decided to run it on a new IP along with my existing pi-hole which was running on a Hyper-V virtual machine. Both have the same uptime, ~24/7.

Being the experimental person I am, I decided to run them both side by side, obviously with different IP's. In my DHCP Server preferences in my Unifi USG, I have set my DHCP nameserver DNS preferences as follows (yes, there are 4 server preferences that can be entered):

DNS 1: IP of pi-hole-1
DNS 2: IP of pi-hole-2
DNS 3: 1.1.1.1
DNS 4: 1.0.0.1

Now for the interesting part; Pi-hole 2 actually receives some lookups as it's logs are slowly becoming populated. pihole-1 certainly captures the vast majority of traffic, however some is being directed to the second pi-hole. This is not a problem per se, however I have a few questions about this. They are as follows:

1. If some DNS lookups are being referred to the second pi-hole server (DNS preference 2), then how much traffic is being sent directly to DNS preference 3, or even 4?
2. Most importantly, what makes a client decide to use a second preference? My understanding was that the next preference would only be used if the previous one was entirely unreachable? Is the preference system more like a load balancer as opposed to a fail-over?

Just looking for someone to demystify this if possible. Thank you for your time!