Summary

As a developer of log collection,
I want to understand how the number of configured kubernetes_sources impacts the kube-apiserver
So that I can understand the limitations with ClusterLogForwarder and make recommendations

Acceptance Criteria

• Documentation (or KCS/Blog) which documents our efforts with an assessment of the results
• Quasi test plan to identify the testing effort

Notes

• We should try to use this information to document best practices and troubleshooting (i.e. limit CLF to 10? container inputs?)

Questions

• Are we able to capture meaning memory and cpu information during testing?
• Can we make any meaningful assessment from monitoring requests from the collector?
• Testing a single container input at collector startup? What metrics can we collect?
• Testing at least three container inputs at collector startup? What metrics can we collect?