NetApp Harvest is an open-source metrics collection and observability platform purpose-built for NetApp storage systems and related infrastructure. Written in Go and released under the Apache 2 license, Harvest functions as an open-metrics endpoint that collects performance, capacity, and hardware metrics from ONTAP, StorageGRID, E-Series storage arrays, and Cisco Nexus switches. The collected data is transformed and routed to time-series databases of the user's choice, enabling comprehensive datacenter monitoring and analytics.
The core functionality centers on gathering detailed metrics from multiple storage and networking platforms and making them available through a standardized metrics interface. Harvest includes pre-built Grafana dashboards that provide datacenter insights without requiring extensive custom configuration. The platform is designed with flexibility in mind, allowing users to collect, augment, and export metrics according to their specific infrastructure needs. New metrics can be added through edits to included template files, making it accessible for users who need to extend monitoring beyond the default offerings.
Harvest's architecture supports integration with Prometheus-compatible time-series databases and Grafana for visualization, positioning it within the standard observability stack used by many organizations. The project includes documentation at netapp.github.io/harvest and offers quickstart guides and Docker Compose examples to facilitate rapid deployment. A notable recent addition is the Model Context Protocol server, which allows users to query Harvest-collected metrics from ONTAP, StorageGRID, E-Series, and Cisco systems through an MCP interface.
The project maintains an active community presence with users engaging through Discord and GitHub discussions.
Harvest addresses a specific gap in NetApp infrastructure monitoring by providing a unified, flexible approach to metrics collection across multiple storage platforms. Rather than requiring separate monitoring solutions for different storage types, organizations can deploy a single Harvest instance to gather metrics from ONTAP systems, StorageGRID deployments, E-Series arrays, and network switches simultaneously. The emphasis on template-based metric definition and the inclusion of pre-built dashboards makes it accessible to organizations without deep expertise in metrics collection and time-series database management.
