RouteSpear documentation

RouteSpear measures the network paths your services actually depend on, from the places that actually use them, and keeps every sample rather than only the averages.

If you are evaluating, start with the quickstart. It takes about ten minutes and ends with real measurements on your screen.

Choose your task

• Investigating a problem now? Start with the symptom and preserve the time range before changing configuration.
• Rolling out monitoring? Follow the deployment journey from one verified agent to a repeatable fleet.
• Preparing a purchase review? Build an evaluation brief with security, deployment and budget evidence.

Follow a complete workflow

• Deploy a fleet with confidence: verify one agent and one measurement, make VDI deployments repeatable, then establish alerts and a budget.
• From symptom to evidence: preserve the incident window, inspect samples and screenshots, correlate infrastructure, and verify recovery.
• Explore the measurement method: work through illustrative examples of silent hops, interrupted uplinks and multiple paths.

What you install

One agent per vantage point. A vantage point is anywhere a measurement is worth taking from: a branch office, a data centre rack, a VM in each cloud region, a laptop at the site where people are complaining.

The agent is a single binary with no runtime dependencies. You download an install pack from the portal, run one script, and it registers itself. There is no database to stand up, no collector tier to size, and nothing to configure before it works.

What you do next

Everything else happens in the portal:

• Create a test: pick what to measure, pick a target, pick which agents run it.
• Read the results: per-hop path views, time series, and the individual records behind them.
• Investigate infrastructure: connect machine identity, device interfaces, host and container resources, passive TCP traffic, and wireless observations.
• Schedule and automate tests: apply running hours, automatic agent assignment, and alert or metric conditions to existing groups.
• Build dashboards: panels you arrange yourself, correlating any metric against any other.
• Query metrics directly: for the question you have not built a panel for yet.
• Set up alerts: health rules, escalation policies and notification channels, so you hear about it instead of finding it on a chart.
• Investigate with Spear Assistant: use Ask Spear beside a page, run skills, or connect an MCP client.
• Monitor services: databases, SIP/RTP, throughput, messaging, directories and file services alongside network diagnostics.
• Monitor a remote host over SSH: Linux resources and an Infrastructure machine, without installing a collector on the target.
• Observe real browser visits: scoped endpoint sessions and page waterfalls, separate from synthetic Browser checks.
• Manage the organization: users and roles, security, secrets, usage, fleet controls and support.

The thing that makes this different

Most network monitoring presents every number as equally true. RouteSpear does not, because measurement is not equally reliable.

A router that rate-limits ICMP, a load balancer that sends consecutive probes down different paths, a Windows host whose operating system will not let the agent control the flow identifier, each of these makes a measurement less certain, and each is recorded on the row rather than smoothed away. When a hop is ambiguous, the interface says the hop is ambiguous.

This matters most during an incident, when the cost of confidently believing a wrong number is highest. See the trust model.

Where to go

If you want to	Read
Get something measuring today	Quickstart
Understand the vocabulary	Core concepts
Know what can be measured	Check types
Get paged when something breaks	Alerts
Run the agent by hand	CLI reference
Work out why an agent is unhappy	Troubleshooting
Run RouteSpear in your own environment	Self-hosting