About Glowo

What we build

Glowo does one job: it tells you that something you are responsible for has stopped working, before a customer does, and it gives you somewhere to point people while you fix it. It checks from the outside, the way your users reach you — not from inside your own infrastructure, which is exactly the thing that may be broken.

Everything in the product exists to serve that job or to keep it from crying wolf. Multi-region checks and confirmation thresholds exist because a single failed request from one place is not an outage. Flap protection exists because a service oscillating between up and down should page you once, not forty times. Maintenance windows exist because the alternative — pausing a monitor before planned work — is how monitors end up paused for six months.

How we decide something is down

A check failing is evidence, not a verdict. Glowo confirms a failure from more than one region and over a configurable number of consecutive checks before it alerts anyone, and it treats timestamps reported by a checker as observations rather than as an ordering key — a machine with a skewed clock should not be able to pin a monitor down or swallow its recovery.

The same conservatism applies to recovery: a monitor comes back up when checks say so, not when a timer expires. A watchdog that changes a status without having checked anything is reporting its own impatience, not your availability.

Where it runs

Glowo is a European company and customer data stays in the EU. Accounts, monitors, check history and status pages are stored on servers in Germany and France. The checker fleet that runs your monitors is distributed globally, so checks happen close to your users, but those checkers hold no customer data — they receive a job and return a result.

The status page you are shown for Glowo itself is not a hardcoded claim of health. It is fetched live from the same public API any customer's status page uses, and when that fetch fails the page says nothing rather than asserting that everything is fine.

Built to be driven by software

Everything the dashboard can do, the API can do. Glowo publishes an OpenAPI 3.1 specification, a Model Context Protocol server for AI agents, and OAuth 2.1 discovery metadata, so an agent can find the surface, request the narrowest useful permissions, and work without anyone pasting a long-lived credential into a config file.

Permissions are one vocabulary across sessions, API keys and agent grants — monitors:read, incidents:write, and so on. A credential also declares where it may be presented, so a key minted for an agent is refused on the REST API and a key minted for CI is refused by the agent transport. Four capabilities — billing, members, workspace settings and the audit log — can never be attached to a credential at all.

How we sell it

The Free plan is not a trial with a clock on it. It has ten monitors, real alerting and a real status page, it needs no card, and it does not expire. Paid plans add regions, faster intervals, longer history, SMS, custom domains and SSO.

There is no sales gate. You sign up, you create monitors, you upgrade if and when you outgrow the plan, and you cancel from the billing page in one click without talking to anyone. Plan limits block new resources rather than generating an overage bill, and existing monitors keep running while you decide. Questions: contact us or email support@glowo.dev.

The company

Glowo is operated by G Networks, registered in France under SIREN 753 647 585, at 100 rue Vulfran Warme, 80000 Amiens, France.

Support and security reports: support@glowo.dev, or the contact form. Live service status: status.glowo.dev. Documentation: docs.glowo.dev.