Products Hosted service

Cloudy Network Monitoring

Run a fleet of programmable probes that continuously check the health of your network.

cloudywithachanceoflatency.net

Cloudy — Run a fleet of programmable probes that continuously check the health of your network.

A hosted monitoring service built around small, independent probes that are deployed and torn down at will, with nothing installed on the hosts running your services and nothing changed in their configuration.

The probes are programmable and remote controlled from a single interface: run this test from A to B, every 5 minutes, for one week, or detect this network condition from all those locations. Tests can involve several probes working in concert.

Measurements keep the packets, not just the metrics: every test yields the actual traffic it generated, centralised and queryable, so a complete picture of a multi-site network can be formed from one place.

Its main access point is an MCP server, so a coding agent can investigate your network on its own.

Cloudy is early and not a commercial product yet. It is looking for engineers willing to try it on a link of their own, and for companies ready to run it on a real industrial site.

Use cases

  • Agentic SRE: let your AI agents run and tailor synthetic network measurements to get exactly what they need;
  • Someone else's network: measure what a link really does before committing to it, and tell whether a slowdown is your software, their IT, or the operator;
  • Cloud migration: measure the real latency cost and check whether performance matches your architecture's assumptions;
  • Online services: measure real network performance end to end, independently of your cloud provider.