Back to all scenarios
Scenario #177
Networking
K8s v1.19, GKE
DNS Latency Due to Overloaded CoreDNS Pods
CoreDNS pods experienced high latency and timeouts due to resource overutilization, causing slow DNS resolution for applications.
Find this helpful?
What Happened
CoreDNS pods were handling a high volume of DNS requests without sufficient resources, leading to increased latency and timeouts.
Diagnosis Steps
- 1Used kubectl top pod to observe high CPU and memory usage on CoreDNS pods.
- 2Checked the DNS query logs and saw long response times.
Root Cause
CoreDNS was under-resourced, and the high DNS traffic caused resource contention.
Fix/Workaround
• Increased CPU and memory limits for CoreDNS pods.
• Enabled horizontal pod autoscaling to dynamically scale CoreDNS based on traffic.Lessons Learned
Proper resource allocation and autoscaling are critical for maintaining DNS performance under load.
How to Avoid
- 1Set up resource limits and autoscaling for CoreDNS pods.
- 2Monitor DNS traffic and resource usage regularly to prevent overloads.