Back to all scenarios
Scenario #190
Networking
K8s v1.19, Azure AKS
Network Latency Between Pods in Different Regions
Pods in different Azure regions experienced high network latency, affecting application performance.
Find this helpful?
What Happened
The Kubernetes cluster spanned multiple Azure regions, but the inter-region networking was not optimized, resulting in significant network latency between pods in different regions.
Diagnosis Steps
- 1Used kubectl exec to measure ping times between pods in different regions and observed high latency.
- 2Inspected Azure network settings and found that there were no specific optimizations in place for inter-region traffic.
Root Cause
Lack of inter-region network optimization and reliance on default settings led to high latency between regions.
Fix/Workaround
• Configured Azure Virtual Network peering with appropriate bandwidth settings.
• Enabled specific network optimizations for inter-region communication.Lessons Learned
When deploying clusters across multiple regions, network latency should be carefully managed and optimized.
How to Avoid
- 1Use region-specific optimizations and peering when deploying multi-region clusters.
- 2Test the network performance before and after cross-region deployments to ensure acceptable latency.