Internet performance monitoring from ThousandEyes, a Cisco company, recorded 459 global outage events during the week of October 19 through October 25, a 19% increase over the prior week's 387 incidents. U.S.-based outages climbed even faster, rising 24% to 293 events, with internet service providers accounting for the bulk of the growth. The data, tracked across ISPs, cloud networks, collaboration platforms, and edge services, points to a volatile stretch for the infrastructure that underpins everyday business connectivity.
ISP Networks Drive the Weekly Spike
ISP outages worldwide jumped 41%, from 138 to 195 events, while U.S. ISP incidents rose 42%, from 77 to 109. This marks the third consecutive week of ISP outage growth, a pattern that suggests carriers are contending with recurring instability rather than isolated incidents. ThousandEyes measures these events using a global network of vantage points that continuously test paths across the internet, flagging when routing, latency, or packet loss crosses thresholds consistent with an outage. Because ISPs sit at the foundation of nearly all internet traffic, disruptions at this layer tend to ripple outward, affecting cloud access, VPN connections, and collaboration tools that enterprises rely on daily.
Public cloud network outages also increased, though more modestly, up 5% globally and 18% in the U.S. Collaboration app networks, by contrast, logged zero outages for the second straight week in both measurements, a reminder that outage patterns vary significantly by service category even when overall totals rise.
A 19-Minute Disruption at a Major Transit Provider
On October 24, PacketFabric - the network transit provider formed through its 2023 merger with Unitas Global - experienced a 19-minute outage beginning around 11:05 p.m. EDT. The disruption initially centered on nodes in New York before spreading to Chicago and London roughly five minutes later. As additional nodes began showing outage conditions, the number of affected downstream regions grew to include Canada, Germany, Taiwan, Ireland, and Switzerland. The incident cleared by 11:25 p.m. EDT.
This pattern - an outage starting at a single geographic node before propagating to other locations - is common in transit-provider disruptions, where routing dependencies mean a localized fault can cascade through interconnected infrastructure before network operators isolate and resolve it.
What the Numbers Mean for IT Teams
For enterprise network administrators, the consistent rise in ISP-level outages over multiple weeks is a more actionable signal than any single incident. Organizations that depend on specific carriers or transit providers for critical paths may want to review redundancy options, since outages at this layer are often outside an enterprise's direct control but still disrupt downstream services. ThousandEyes' outage counts reflect detected anomalies across monitored paths rather than a complete census of every internet disruption, so the figures are best read as directional indicators of relative network health rather than precise measures of total incidents.