MAAS' DNS fails to reply while deploying

[Tony]

This didn’t happen with MAAS 2.6, but now that we’ve upgraded to 3.3.5, we keep getting DNS query drops whenever we deploy a host.
I made a script that flushes my DNS caches and every 10 seconds queries for an entry on our MAAS server. On Wednesday I deployed hosts, and just a few minutes ago I began deploying. The log looks like this:

Wed Feb 28 01:57:26 PM UTC 2024: DNS lookup failed!
Wed Feb 28 01:57:36 PM UTC 2024: DNS lookup failed!
Wed Feb 28 02:01:17 PM UTC 2024: DNS lookup failed!
Fri Mar 1 06:22:15 AM UTC 2024: DNS lookup failed!
Fri Mar 1 06:23:45 AM UTC 2024: DNS lookup failed!
Fri Mar 1 06:23:55 AM UTC 2024: DNS lookup failed!
Fri Mar 1 06:24:05 AM UTC 2024: DNS lookup failed!

And the failures just keep on coming until the deployment is complete. Then not a single failure.

Naturally our cloud-init scripts at the deployment phase will also fail the DNS lookups and thus fail the deployment.

The loads on the servers is negligible, and the server (VM) has 4 vCPUs and 8 GB or RAM. It should easily handle this.

This script is a simple bash script (here with obfuscated info):
#!/bin/bash

while true
do
  resolvectl flush-caches
  if ! nslookup nfs01.ourdomain.com <ip of maas> >/dev/null 2>&1; then
    echo "$(date): DNS lookup failed!"
  fi
  sleep 10
done