Static Web App in West Europe returned HTTP.sys 503 "Service Unavailable" for ~6 minutes on 18 Aug 2026 (and 12 Aug) — was there a platform event?

Our production Azure Static Web App (Standard plan, West Europe, no linked Functions/API) became unavailable for ~6 minutes on 18 August 2026, from approximately 07:09 to 07:15 UTC. We have strong evidence this was a platform event rather than anything on our side, and we'd like to know whether Microsoft can confirm an incident, and what customers should do in this situation.

What we observed

Two independent witnesses failed simultaneously against the two different scale units serving our app's partition (azurestaticapps1.trafficmanager.net):

1. Our external uptime monitor requests the *.azurestaticapps.net URL directly.
   It resolved to 51.124.12.35 (waws-prod-am2-8b19502a, slice msha-slice-1-am2-1) and received a fast, well-formed HTTP 503 with healthy timings (DNS 16 ms / TCP 23 ms / TLS 155 ms / server_time 150 ms). The response headers were:

      Date: Tue, 18 Aug 2026 07:10:06 GMT
      Content-Length: 326
      Content-Type: text/html; charset=us-ascii
   

   i.e. the default HTTP.sys "Service Unavailable" response — a kernel-level
   rejection on the scale-unit front end, generated before any application
   layer runs (which is also why there is no x-azure-ref to share).
2. At the same time, our own reverse proxy (separate network path), which was pinned by DNS to the other slice — 20.50.153.39 (waws-prod-am2-5ecab9f3, slice msha-slice-1-am2-0) — logged repeated "upstream timed out (110) while reading response header" errors from 07:09:34 to 07:14:59 UTC, plus fast 503s proxied through from that slice.

Supporting signals

• The resource's SiteHits metric collapsed to ~0 in the 07:13 UTC bucket and jumped back to normal (~190/bucket) right after; SiteErrors recorded no data, consistent with the 503s being generated upstream of the telemetry layer.
• Activity Log shows no deployments, configuration changes, or administrative operations on the static site that morning.
• No Resource Health or Service Health event was raised for the resource, and nothing appears on the public status page.
• It recovered on its own (~07:15 UTC) with no deployment on our side.
• It appears to be recurring: the same monitor stored an identical HTTP.sys 503 sample dated 12 Aug 2026 06:50:30 GMT — six days earlier, nearly the same time of day.

Questions

1. Can anyone from the Static Web Apps team confirm whether there was a platform incident affecting SWA partition azurestaticapps1 / scale units waws-prod-am2-5ecab9f3 and waws-prod-am2-8b19502a in West Europe on 18 Aug 2026 ~07:09–07:15 UTC (and possibly 12 Aug ~06:50 UTC)?
2. Why would such an event not surface in Resource Health for the affected resource? Is there any signal (metric, log, health event) that reliably captures platform-level 503s for a functionless Static Web App, other than enabling diagnostic logs going forward?
3. Our subscription has no technical support plan, so we cannot open a support case. What is the recommended path for getting platform incidents like this acknowledged or investigated in that situation?

Thanks in advance.