Back to all scenarios
Scenario #172
Networking
K8s v1.20, On-premise
Network Latency Caused by Overloaded VPN Tunnel
Network latency increased due to an overloaded VPN tunnel between the Kubernetes cluster and an on-premise data center.
Find this helpful?
What Happened
The VPN tunnel between the Kubernetes cluster in the cloud and an on-premise data center became overloaded, causing increased latency for communication between services located in the two environments.
Diagnosis Steps
- 1Used kubectl exec to measure response times between pods and services in the on-premise data center.
- 2Monitored VPN tunnel usage and found it was reaching its throughput limits during peak hours.
Root Cause
The VPN tunnel was not sized correctly to handle the required traffic between the cloud and on-premise environments.
Fix/Workaround
• Upgraded the VPN tunnel to a higher bandwidth option.
• Optimized the data flow by reducing unnecessary traffic over the tunnel.Lessons Learned
Ensure that hybrid network connections like VPNs are appropriately sized and optimized for traffic.
How to Avoid
- 1Test VPN tunnels with real traffic before moving to production.
- 2Monitor tunnel utilization and upgrade bandwidth as needed.