Check Prometheus metrics of coredns pod and determine health/issues with DNS in EKS cluster

[joshisumit]

CoreDNS exports basic process and Go runtime metrics as well as CoreDNS-specific metrics, on port 9153/metrics

https://coredns.io/plugins/metrics/

Get metrics from CoreDNS in real time to visualize and monitor DNS failures and cache hits/misses.

Important Metrics:

1. coredns_dns_request_count_total - will show you how busy CoreDNS is, and you can look deeper to understand how requests are being resolved.

# HELP coredns_dns_request_count_total Counter of DNS requests made per zone, protocol and family.
# TYPE coredns_dns_request_count_total counter
coredns_dns_request_count_total{family="1",proto="udp",server="dns://:53",zone="."} 957

2. coredns_dns_request_duration_seconds - metric can show you how much DNS latency is contributing to overall, user-facing response time.

# HELP coredns_dns_request_duration_seconds Histogram of the time (in seconds) each request took.
# TYPE coredns_dns_request_duration_seconds histogram
coredns_dns_request_duration_seconds_sum{server="dns://:53",zone="."} 0.6624103219999997
coredns_dns_request_duration_seconds_count{server="dns://:53",zone="."} 957

3. coredns_dns_response_rcode_count_total - Check error codes of DNS responses
   Errors like NXDomain and FormErr can reveal a problem with the requests CoreDNS is receiving, while a ServFail error could indicate an issue with the function of the CoreDNS server itself.

# HELP coredns_dns_response_rcode_count_total Counter of response status codes.
# TYPE coredns_dns_response_rcode_count_total counter
coredns_dns_response_rcode_count_total{rcode="NOERROR",server="dns://:53",zone="."} 521
coredns_dns_response_rcode_count_total{rcode="NXDOMAIN",server="dns://:53",zone="."} 436

[github-actions]

Hello @joshisumit , thank you for submitting an issue!