Otel perf test

[Paamicky]

Description of the issue

The eks_daemon:otel_performance_test (test/otel/performance) was failing with no data returned for k8s.pod.cpu.utilization / no CPU data. The OTel kubeletstats receiver deprecated k8s.pod.cpu.utilization and replaced it with k8s.pod.cpu.usage, so the test's query for the old metric name returned empty. Separately, the DaemonSet CPU threshold was calibrated when the test was first added and sits just below the agent's actual steady‑state usage.

Description of changes

• Query k8s.pod.cpu.usage instead of the removed k8s.pod.cpu.utilization (query, case branches, and performance_thresholds.json). No unit change, the value is still CPU cores, so the existing % of node math, thresholds, and stored regression baselines remain valid.
• Bump the DaemonSet CPU threshold 1.30% → 1.50% to match measured steady‑state usage (~1.5% of node CPU, confirmed against live Container Insights data, which shows CPU has been flat).

License

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

Tests

• go vet -tags integration ./test/otel/performance/ and gofmt clean.
• Verified via Query Studio / CloudWatch Container Insights that k8s.pod.cpu.usage is emitted for the agent pods