Skip to content



Orb is a new kind of observability platform, fully open source and backed by NS1 that helps operators, developers, and end users understand their networks, distributed applications, and traffic flows in real time.

It combines concepts from Internet of Things (IoT), Edge Compute, and high throughput stream processing to offer command of a fleet of distributed agents which can be orchestrated in real time to summarize high-volume, information-dense data streams down to lightweight, actionable data sets directly at the edge.

Orb gives you a precision instrument to collect just the observability data you are interested in, extracting the needles from the haystack at the edge. It is based on the pktvisor observability agent.

The platform is open-source, extensible, vendor-neutral, and cloud-native.

Star on GitHub


Born at NS1 Labs, Orb is a new kind of observability platform, fully open source and backed by NS1. Orb orchestrates network observability policies across a fleet of agents on the edge in real time, allowing you to dynamically filter and collect precise “small data” from your global network, providing you with lightweight, immediately actionable results.

How is Orb Different?

Orb focuses on edge analysis, preferring “small data” style actionable metrics over collection and storage of terabytes of raw, inscrutable data: think of it as a scalpel rather than a machete.

  • Powerful control plane to adjust analysis and collection parameters in real time across the entire fleet
  • Designed to plug in to popular observability stacks, like Prometheus and Elasticsearch, plus cloud storage and data pipelines
  • Visualize and automate on results both at the edge (for hyper-real-time local view) or centrally in the cloud (for global view)
  • Fast, reliable, secure, modular, extensible, open-source

How Orb Works

Orb is composed of a side-car style agent (the open-source pktvisor) that performs edge analysis on network data streams, combined with an IoT-inspired central control plane and UI that provide fleet and configuration management functionality.


  • Resource-efficient agent that analyzes data streams in real time
  • Dynamic administration and configuration
  • Modular system of inputs and analysers
  • Cloud-native, microservices-based architecture
  • Self-host (docker-compose or k8s) or SaaS option
  • Manages fleet of pktvisor agents, including heartbeats, tagging and grouping
  • Orchestrates dataset policies to decide what data to extract from which agents
  • Central collection and exporting of summarized results that are then available for analytics, security, automation, etc.
  • Provides a single pane of glass across all sensors
  • Provides a deep dive into visualizations and analytics that can help you make more effective network decisions in the future