Retina is a cloud-agnostic, open-source Kubernetes network observability platform developed by Microsoft that provides centralized monitoring for application health, network health, and security. Written primarily in Go, it leverages eBPF technology to collect customizable telemetry from Kubernetes workloads without requiring kernel modifications. The platform exports metrics to multiple storage backends including Prometheus and Azure Monitor, and supports visualization through Grafana, Azure Log Analytics, and other vendor tools, making it suitable for cluster network administrators, security administrators, and DevOps engineers managing DevOps, SecOps, and compliance requirements.
The platform offers two major feature sets: Metrics and Captures. The Metrics capability provides industry-standard Prometheus metrics for continuous cluster monitoring, with both Basic and Advanced modes available. Advanced metrics include packet parsing functionality but may experience performance considerations on high-core-count systems with 32 or more CPU cores under heavy network load. The Captures feature enables on-demand packet capture for deep network investigation, accessible through both a CLI tool installable via Krew and a Kubernetes Custom Resource Definition approach. Retina supports multiple operating systems including Linux, Windows Server 2022, and Azure Linux, though Windows Server 2019 is no longer supported.
Installation is streamlined through Helm charts distributed via GHCR, with published images cryptographically signed for provenance verification using cosign. The platform emphasizes configurability and actionable insights, allowing users to investigate network issues on-demand while maintaining continuous monitoring capabilities across their clusters.
The project follows Microsoft's open-source governance model, requiring Contributor License Agreements and adhering to the Microsoft Open Source Code of Conduct. Community meetings are hosted periodically to facilitate discussion and contribution guidance, with detailed documentation available at retina.sh covering installation, configuration, and contribution procedures.
