[Feature] Container Network Metrics Filtering

[msftjonw]

Customers often face challenges with excessive metrics ingestion in containerized environments, leading to high storage costs and noisy dashboards. This new feature allows customers to define precise filters based on namespace or pod at the Cilium level using a CRD, ensuring only relevant metrics are collected before reaching Prometheus. The result is reduced data volume, lower operational costs, and improved observability efficiency.

ETA: Public preview in November, 2025