This Container does provide an option to measure DNS latency better and does expose the following metric:
dns_latency_millisecondsdns_failure_failure_total
A potential query rule to validate a problem could be increase(dns_failure_failure_total[10m]). If there is a spike in failures, it's definitely worth further investigating.