Back to all scenarios
Scenario #153
Networking
K8s v1.20, on-premise

Network Performance Degradation Due to Misconfigured MTU

Network performance degraded due to an incorrect Maximum Transmission Unit (MTU) setting.

Find this helpful?
What Happened

Network performance between pods degraded after a change in the MTU settings in the CNI plugin.

Diagnosis Steps
  • 1Used ping tests to diagnose high latency and packet drops between nodes.
  • 2Verified MTU settings on the nodes and CNI plugin, and found that the MTU was mismatched between the nodes and the CNI.
Root Cause

MTU mismatch between Kubernetes nodes and the CNI plugin.

Fix/Workaround
• Aligned the MTU settings between the CNI plugin and the Kubernetes nodes.
• Rebooted affected nodes to apply the configuration changes.
Lessons Learned

Ensure that MTU settings are consistent across the network stack to avoid performance degradation.

How to Avoid
  • 1Implement monitoring and alerting for MTU mismatches.
  • 2Validate network configurations before applying changes to the CNI plugin.
Previous ScenarioNext Scenario