FAQ

Why does this number differ from what atlas.ripe.net shows?

Usually one of three things:

1. Freshness. Lookups in Atlas are live, but aggregates read an archive with its own cadence — see Data sources. A fleet count can be up to a day old by design.
2. Time zone. Everything in Atlas is UTC. RIPE's interface may render in your local zone, which shifts day boundaries.
3. Cohort. Two views can count subtly different things — "measurements that started today" is not "measurements running today". See Trends.

Why does a chart stop before today?

Because most series are bucketed by UTC day and built from archived data, so today is incomplete or not yet imported. That is the normal state of a daily series, not a fault. See Data delay.

Why is there a hole in the middle of a series?

A missed ingestion window. Those can be backfilled after the fact, so a gap you see today may be filled tomorrow. A gap is not a measured zero — see Gaps versus zeroes.

Can I run measurements from Atlas?

No. Atlas is a read surface. The exceptions are your dashboard, which can stop a measurement you own, and monitoring checks of the UDM kind, which do create measurements on RIPE and do spend your credits.

Why is my probe in the wrong city?

Probe locations are declared by the host and are not verified. Atlas cross-references them against geofeeds, commercial databases and latency constraints, but it does not overwrite RIPE's record. If it is wrong, raise an issue against that probe — that is exactly what flags are for. See Geofeeds and geolocation.

Why is a prefix marked anycast when I know it isn't?

The anycast data is imported weekly from manycast.net and is probabilistic rather than authoritative. Check the per-protocol confidence scores and the partial flag before treating it as fact — a partial prefix is only anycast in part. See Anycast.

Conversely, a genuinely anycast prefix may be missing simply because the upstream hasn't observed it yet.

How do I get staff access?

It is assigned, not requested through the UI. Most of the analysis-heavy views — health, mesh, deployments, monitoring, PeeringDB, anycast — are staff-gated, so if the site looks smaller than this documentation describes, that is why. See Accounts and roles.

I set my dark-mode preference on the Settings page and nothing happened.

The switches on /settings are placeholders that aren't wired to anything. The real theme toggle is in the top app bar. See Profile and settings.

Is there an API?

Yes — see the API reference. It is generated from the service's own OpenAPI description, and the full specification is downloadable at /openapi.json. Personal API keys require a staff account.

Why do I need an API key at all if the data is public?

You don't, for public endpoints. Keys are for account-scoped and role-gated endpoints. See Authentication.

I lost my API key secret. Can I get it back?

No — only a hash is stored. Issue a new key and revoke the old one; see Rotating without downtime.

I forgot my password and never set an email address.

There is no recovery path without an email address, since reset works by sending you a link. If you still have a session, set an address on your profile now.

How do I report a problem?

Two different things:

• The data is wrong → raise an issue against the record.
• The site is wrong → send feedback, signed in if you want a reply.

Something isn't answered here

Send feedback — questions that come up more than once end up on this page.