KASPR Datahaus · 15-minute record
The Iberian blackout of 28 April 2025, seen from Lisbon at 15-minute resolution
Window: 27–29 April 2025 · Lisbon, Portugal · 15-minute export
Shortly after half past eleven on the morning of 28 April 2025, the power grid across Spain and Portugal collapsed. Lisbon’s internet did not go dark all at once: the addresses answering our probes halved within a quarter hour, then kept sliding through the afternoon to 27% of normal, and did not recover until around midnight, 12.75 hours after the first drop. This page sets out that record and offers the rows for download.
1,857
addresses answering per quarter hour, normally
27%
of normal at the lowest point
5
hours below half of normal
12.75
hours from first drop to recovery
What happened
The grid went first
At 12:33 Central European time, 11:33 in Lisbon, the Iberian electricity grid lost most of its generation within seconds, and Spain, Portugal and parts of south-western France lost power. Restoration ran through the afternoon and evening, area by area, and was largely complete in Portugal by the early hours of 29 April. The internet did not fail on its own; it followed the power.
The measurement
Half at once, then a long slide
In the 24 hours before the blackout, about 1,857 addresses in Lisbon answered our probes each quarter hour. At 11:30 local time, the quarter hour containing the grid collapse, the count fell below 90% of normal, and within the next quarter hour it had halved. Unlike a switch-off it then kept falling: the lowest point came at 18:45, at 503, 27% of normal. It stayed below half of normal for 5 hours and was back above 90% by 00:15.
Times in the text are local (Europe/Lisbon); the chart axes are UTC.
The adjusted responding count: distinct addresses in the district answering our probes, on the export’s reference scale. The shaded span runs from the first quarter hour below 90% of normal to the first hour back above it. Axis in UTC.
A flat series of responding addresses that drops sharply on the event day and returns.Latency and vantage points
Slower answers from the addresses that stayed up
Round-trip time rose from a median of 192 ms before the blackout to 205 ms during it. Scanner coverage held: at least 6 of 7 paths saw responses in every quarter hour, so the count was thinning at the destination, not at the vantage points.
Mean round-trip time per quarter hour, in ms, across every response in the bucket. Fewer, farther or busier respondents move this line even when the count does not. Axis in UTC.
A latency series that rises during the event window and settles afterwards.The shape
What a blackout looks like in measurement
The shape is the opposite of a shutdown. The first drop is sharp, as the addresses with no backup power vanish with the grid, but the decline then continues for hours, consistent with routers, exchanges and towers running on batteries or generators dropping out one by one, and the recovery is gradual as power returns area by area. Read together with the Dhaka record, the two events show why the 15-minute resolution matters: at daily resolution both are simply a bad day.
As with every event on these pages, the measurement records when, how deep and in what shape; the cause comes from the public record.
Take the data
Download the sample behind this page
Three days at 15-minute resolution, the day before, the day of and the day after, the same window the charts above cover: round-trip time and its variance per quarter hour, the scanner paths and cohorts that saw a response, and a provenance label for every bucket, with the district named by country and administrative level. The responding counts and the expected-schedule and calibration columns stay in-house; the codebook carries their caveats.
How to read it
Every quarter hour, scanners outside the country probe a fixed sample of addresses in the district. The count is how many answered, rebased so that months with different scanner coverage compare; it is a relative measure, not a population. Normal is the median of the 24 hours before the documented onset. The measurement records timing, depth and shape; the cause comes from the public record.
Independent, coarser corroboration: Cloudflare Radar outage centre
The same 15-minute measurement, replayed for another district: Open Live