Documentation menu
Self-hosting
Running RouteSpear inside your own environment, what it involves, and when it is the right choice.
RouteSpear can run entirely inside your environment, with no measurement data leaving it. This is arranged per deployment rather than self-served, so it starts with a conversation.
Talk to us about self-hosting.
When it is the right choice
Self-hosting exists for environments where the cloud service is not an option rather than not a preference:
- Segregated or air-gapped networks: defence, government, utilities and industrial control environments where telemetry cannot leave.
- Regulatory constraints on where measurement data may reside.
- Security review outcomes where a continuously-updated map of internal network topology is judged too sensitive to hold externally, which is a reasonable conclusion and one we would rather discuss than argue with.
If your reason is preference rather than constraint, the hosted service is almost certainly the better outcome. It is cheaper, it upgrades itself, and the time your team spends running a measurement platform is time not spent using one.
What it involves
A self-hosted deployment is a real system, not a single binary. It includes the ingest and processing services, a time-series store for measurements, a relational store for configuration, and a message queue between them.
We deploy and support a defined configuration rather than handing over a set of containers, because the difference between those two is the difference between something that works in year two and something that quietly diverges from every other installation. That means:
- A certified deployment, sized for your agent count.
- One supported release train, upgraded on a defined schedule.
- A named support contact, and a diagnostics bundle so problems can be investigated without direct access to your environment.
The middle option
Most requests for self-hosting are really about where the measurement data lives, not about who operates the web interface.
If that describes your situation, ask about a customer-hosted data plane: the collection and storage run inside your boundary while the portal stays hosted. It answers the data-residency question without you taking on the operation of a database cluster, and it is a much faster path to running.
Agents are the same either way
The agent, the check types, the portal and the measurements are identical in both models. Choosing one does not mean a different product or a reduced feature set, and it is not a decision that has to be made before you have evaluated anything, start on the hosted service and talk to us in parallel.