Major outages and post-mortemsBackground
Google Cloud outage on June 12, 2025 disrupts Workspace and apps built on it
From 10:51 a.m. to about 6:18 p.m. PDT on June 12, 2025, a crash in Google Cloud’s Service Control system caused 503 errors across many Google Cloud products and Google Workspace. The same afternoon, Cloudflare said a third-party cloud provider’s storage failure broke WARP connections and Access logins for about two and a half hours. For a remote worker this looked like “my internet is down” even though home connections were fine. Published October 6, 2026.
- Event date
- Published here
- Updated
- Source
- Google Cloud Service Health incident report
Background itemThis happened 481 days before we published it (more than 90). It is here as context for the guides it links, not as breaking news. Check the source for anything that has changed since.
What changed
On the afternoon of June 12, 2025, Google Cloud had a large outage. Google’s incident report says many Google Cloud products returned errors, and that Google Workspace products saw increased 503 errors in external API requests. Apps built on Google Cloud failed for their users at the same time.
Google traced the failure to Service Control, a system that checks API requests. A policy update with unintended blank fields hit a code path without proper error handling, and the resulting null pointer crashed the binary. Google said the new check was not protected by a feature flag, so the problem spread globally instead of being caught in one region.
The same afternoon, Cloudflare said a failure in a third-party cloud provider’s storage broke its Workers KV service. That in turn broke Cloudflare WARP new connections and Access identity logins, which some employers use for remote access, from 17:52 to 20:28 UTC.
When
- Google Cloud: 10:51 a.m. to about 6:18 p.m. PDT on June 12, 2025 (1:51 p.m. to 9:18 p.m. EDT).
- Cloudflare WARP and Access: 1:52 p.m. to 4:28 p.m. EDT, about two and a half hours.
Who is affected and what it means for you
For a remote worker, this kind of outage feels like “my internet is down”: Gmail, Drive or a company app stops loading, a remote login fails, and the natural move is to restart the router. But the home connection was fine. A Starlink line, a cable line and a cellular hotspot all reached the same broken service.
That matters for planning. A backup internet line protects you from a failure on your path: the dish, the satellite network, your local carrier. It does nothing when the service at the other end is down. People who switched to a hotspot during this outage spent mobile data and gained nothing.
It also matters for troubleshooting. If you run a VPN or a zero-trust client like WARP, a provider failure can look exactly like a VPN problem caused by CGNAT. Our VPN guide covers the real CGNAT issues; this was not one of them.
What you can do now
- Before you restart anything, check two status pages: the app’s (for example the Google Cloud or Google Workspace status page) and a general site you trust. If other sites load fine, your connection is probably not the problem.
- Check your own link quickly. The Starlink app’s statistics page shows outages and latency on your side. If it shows no drops, the problem is likely upstream.
- Keep one non-cloud fallback for urgent work: a phone number for your team lead, a calendar copy on your device, documents you need today stored offline.
- Our troubleshooting page lists the failures that are on your side and how to tell them apart.
What stays the same
- Nothing about Starlink changed. This was not a Starlink outage.
- The advice for a backup line is unchanged: it covers failures on your path, not failures at the service.
What we don’t know yet
- Google’s report does not give a count of affected users.
- Cloudflare’s post names a third-party cloud provider without, in the part we read, giving a separate timeline for that provider’s failure. We have not tied the two together beyond what each company wrote.
- We have no logged data from June 2025; the owner’s logger had not started.
This item is background: it happened more than 90 days before we published this page.
Source
Multiple GCP products are experiencing Service issues (Google Cloud Service Health incident report)
Also cited
Related guides on this site
Correction log
No corrections since publication. Spotted an error? See corrections for how to tell us; changes are logged here with the date.