Who went down, and why

From each vendor's own public status page (Atlassian Statuspage and incident.io JSON, Google Cloud's incidents feed, AWS Health history, Slack's status API), the postmortems vendors publish, and SEC Form 8-K Item 1.05 filings for material cybersecurity incidents. Times are as the vendor reported them, converted to UTC.

Every day at 03:30 UTC. Status feeds are read in full each night; older history is filled in a few vendors at a time.

Only from the vendor’s own words. A rule reads the sentences where the vendor explains why ("caused by", "due to", "root cause") and maps them to one of fourteen categories. When the vendor has not said, the cause is "Not disclosed". Landmark postmortems are categorized by hand.

The vendor’s own impact rating. On Statuspage pages, minor is degraded performance or a partial problem, major a partial outage, critical a full outage. Google Cloud’s low, medium and high and AWS’s informational, degraded and disruption levels are mapped to the same three.

No. This is what vendors chose to post. Some post every blip, some only large outages, so compare a vendor with itself over time rather than one vendor with another. For promised versus measured availability, see slas.fru.dev.

Since December 18, 2023, US public companies must disclose material cybersecurity incidents on Form 8-K, Item 1.05, within four business days of deciding they are material. Those filings are the most reliable public record of corporate breaches.