Outages and root causes.

Two years of public outages, and what each company said caused them. Every entry links to the company's own write-up; nothing here is our speculation.

Status pages re-read 32 min ago

Incidents with a published root cause

By month, last two years · Google Cloud

Peak 0 · December 2024

Incidents listed: 0; With a published RCA
Official write-ups: 0; Linked from the incident
Major outages: 0; As graded by each company
Services: 0; With at least one entry

Notable outages, summarised

Read from each company's published analysis and summarised in our own words. Follow the link for the full account.

Google Cloud
Oct 20, 2026
Power loss in europe-west2 (London) zone
A London zone lost power, taking down zonal VMs, disks, databases and clusters and causing elevated errors across the europe-west2 region for about six hours (starting late on 19 October Pacific time).
Root cause, as stated by Google Cloud
A utility cable fault cut one power feed and misconfigured generator control logic failed to start the backup generators, so the zone went dark once the UPS batteries ran out.
Read the official analysis (opens in a new tab)Report an error (opens a correction request in a new tab)
Google Cloud
Aug 20, 2026
Multi-product degradation in us-west1 (Oregon)
Google lists 27 products in us-west1, including Compute Engine, GKE, Cloud Run, Cloud Storage and IAM, that had elevated latency, errors and provisioning failures for a little over two hours, with some products taking longer to clear backlogs.
Root cause, as stated by Google Cloud
Google says planned fiber maintenance unexpectedly cut capacity between data centers and automated rerouting failed, congesting the network and sharply slowing core metadata and database systems that other products depend on.
Read the official analysis (opens in a new tab)Report an error (opens a correction request in a new tab)
Google Cloud
Jun 12, 2025
Global API outage from Service Control crash loop
API requests to dozens of Google Cloud products, and several Google Workspace services, failed with 503 errors worldwide; most regions recovered within about two hours while us-central1 took close to three, and the incident was closed after about seven and a half hours once lingering product issues cleared.
Root cause, as stated by Google Cloud
A policy change containing blank fields hit new Service Control code that had no error handling or feature-flag protection, causing a null-pointer crash loop in every region.
Read the official analysis (opens in a new tab)Report an error (opens a correction request in a new tab)

From the status pages

With a published RCA Partial and major outages Both

Showing incidents whose vendor named ap-southeast-2. Incidents with no stated region are not included, though they may have affected it too.

Nothing matches these filters.

Who publishes root causes

Write-ups posted on status pages, last two years

Public information only

Everything on this page was already published by the companies themselves or by public bodies. statuswall gathers it in one place, links every item to its original, and adds nothing from private or leaked sources.

Sources

Official status pages: incident titles, severity, times and the link to each published root-cause analysis, exactly as each company posted them
Company engineering blogs and incident reports: the basis for the notable-outage summaries, each linked from its card

How fresh it is

Incidents and RCA links: re-read from every status page about once an hour; last run 32 min ago
Notable-outage summaries: written by hand and last checked against their sources on Nov 10, 2026

Titles, severity and times are as published on each company's status page. "Published RCA" means the company posted a postmortem on that incident's official page; we link to it and do not reproduce it. Many companies publish root causes on a blog instead, or not at all, so the absence of a link is not evidence that no analysis exists. Severity grading differs between companies and is not comparable across them. This is not a complete record of every outage.

statuswall is independent and is not affiliated with, sponsored by or endorsed by any company named here. Company and product names are used only to identify whose public information is being referred to. The information is provided as is, for general reference, with no guarantee that it is complete, current or free of error; it is not security, legal or purchasing advice. Where this page and an official source differ, the official source is right.

Outages and root causes.

Notable outages, summarised

Power loss in europe-west2 (London) zone

Multi-product degradation in us-west1 (Oregon)

Global API outage from Service Control crash loop

From the status pages

Public information only

Sources

How fresh it is