Tutorial · data freshness
Public /status groups data by purpose — Ukraine prices, EU markets, grid operations, forecasts, reference data — and shows each group's freshness state and data age. When a group is not fresh, the page usually names the source that caused it.
/status shows data groups as fresh, stale or pending.
Each group shows how current its data is and how old it is; data that already covers tomorrow's delivery day is marked separately.
Only the Alvo team sees the detailed source SLAs.
On /status, data is grouped by purpose: Ukraine prices, EU markets, grid operations, forecasts, reference data. Start with the group your signal relies on.
Fresh means the group's data is within its freshness limit; stale means at least one source in the group has gone past it; pending means the first observation has not been published yet.
When a group is not fresh, the source that set its state usually appears under its name, followed by how current the data is and how old it is. The Forecasts group names no source, and neither does a reading marked 'last known'. A 'Recent capture delay' or 'Prolonged capture gap' label shows how long new data has been missing.
When a group behind a forecast, brief or export is stale, read them with that caveat, not as fresh data.
A fresh ENTSO-E mark means the fetch attempt was archived on time; it does not prove the response held usable market data.
The archive records each forecast before the delivery window closes; a missed day is not rebuilt afterwards.
When Alvo sees a gap, the affected group on the public status page stops reading as current.
The page groups data by purpose and, where it can, names the source that set a group's state. The Forecasts group names no source, and neither does a reading marked 'last known'.
Not always. When an upstream source returned no data for a whole day, the forecast for those delivery days cannot be recorded after the fact: the forecast archive keeps the gap. Capture usually carries on the next day: when the main model lacks a history day it needs, a simpler fallback model produces the forecast. Only the delivery days that rest on the missing day itself are lost, and the freshness state turns fresh again with the next capture.
Open /status: service and data-group states are visible without signing in.
