Palo Alto Networks had three separate incidents on 20 October 2026, two of them tied to WildFire sample processing in UK and Germany cloud, and one affecting Cortex XDR, XSIAM, XSOAR 8, XPANSE and Cloud in the UK region for nearly 18 hours. The underlying fault traces back to Palo Alto Networks' cloud vendor, not to anything in your own setup.
Started
07:52 UTC
Duration
Lasted 17h 57m
Source
IsDown
Next time
Palo Alto Networks is degraded. Is your product working?
An incident on their side does not always mean an outage on yours, and the only way to know is to already be watching your own endpoints. Set that up in under a minute and you will be notified the next time your own site is affected. And there will be a next time.
50 monitors for your own websites. No card.
Affected components: United Kingdom, WildFire UK Cloud
What happened?
The first incident started at 06:52 UTC, when Palo Alto Networks reported WildFire experiencing an issue in the UK Cloud, with customer impact described as delay in sample processing and client disconnection errors. By 07:37 the company said the cause was an outage at GCP, and updates through the day repeatedly pointed back to GCP's own status page for the europe-west2 region. That WildFire incident ran until it was marked resolved at 16:15, a duration of 9h 24m. A second, separate incident began at 07:52 UTC affecting Cortex XDR, XSIAM, XSOAR 8, XPANSE and Cloud, described as degraded performance in the UK region. Palo Alto Networks said the root cause had been identified quickly, then at 10:30 attributed the issue to its cloud vendor as well. Updates at 11:50 and 14:17 described gradual improvement, a fix was implemented at 16:06, and the incident was not marked resolved until 01:44, giving a total duration of 17h 57m, the longest of the day. A third, shorter WildFire issue hit the Germany Cloud at 11:28, with frequent disconnection and delay in sample processing reported, and was resolved by 12:24 after 56 minutes. Three users in the United Kingdom filed reports during the window: at 07:22 one wrote that "Globalprotect is connecting but not providing connectivity", at 08:21 another said "GlobalProtect is not working", and at 10:15 a third reported the "Global protect VPN not able to connect." None of these reports map directly to a GlobalProtect-specific incident filed by Palo Alto Networks itself, though the timing overlaps with the UK region degradation. Across 90 days, Palo Alto Networks has logged 24 incidents with a median duration of 4 hours 12 minutes, so a near-18-hour incident sits well outside the usual pattern.
Learning
UptimeRobot's own probes, watching Prisma across EU and NA regions on a 60 second interval, stayed at 100% uptime over 24 hours and above 99.9% over 7 and 30 days, and this incident was not registered against those checks. That is not a contradiction. The probes watch a specific component you point them at, and the failures on this day sat in WildFire sample processing and in Cortex's backend services, both of which depend on processing capacity supplied by a third-party cloud vendor rather than on the reachability of an endpoint you'd normally ping. A check that only confirms your login page loads or your API responds to a handshake will not see degraded sample processing or slow backend throughput sitting behind it. The GlobalProtect connection reports from UK users line up with that pattern too, since a VPN that connects but provides no onward connectivity often passes a simple reachability check while failing the thing people actually need. Your own monitoring covers your half of the stack and will tell you if your product is reachable, but it cannot see inside Palo Alto Networks' backend or its cloud vendor's region. UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.