[TRACKING] Scalability Test Follow-Ups #409

Summary / Objective

Though #284 is in a good shape for merging, there are a few items that need to be worked on next.

Checklist / Sub-tasks

Fix:

• Remove usage of pkill in the scale test spec
• Remove hardcoded timeouts
• Add configuration for Gomega default timeout

Improvements:

• Add kind runtime for in-cluster testing
• Add more test scenarios (node count, rule count, enforcement modes, flapping conditions)

Assess:

• Evaluate scraped metrics (provided value/insights)

Experimental:

• Add JUnit report for graph view in TestGrid
• Add a performance dashboard

CI:

• Add the scale test as presubmit ProwJob
• Add canary job using KWOK's kind runtime
• Coordinate with NPD for scale test in a real 100-node cluster

Design Documents & References

LFX sync on August 13, 2026: https://docs.google.com/document/d/1xBTqF8LaKfUrqb5AODDOVRMBlf9e5Hwsb0znW-rkLVg/edit?tab=t.0#heading=h.3qgt4dwa9a2l

Target Milestone / Release

Next