10 causes of cloud outages, from config errors to heat waves

A cloud outage occurs when a fault in shared infrastructure, such as networks, identity systems, or data centers, disrupts services like email or chat, even if the core application itself is intact. Real incidents at Microsoft, AWS, Google, and Meta illustrate ten distinct causes, from configuration errors to heat waves. A single faulty configuration change on Microsoft's Azure network disrupted Outlook and Teams in January 2023, while a defective CrowdStrike update crashed 8.5 million Windows devices in July 2024. Expired certificates blocked Teams sign-ins in 2020, and Meta's BGP routing failure took its services offline in 2021. DDoS attacks, sign-in key errors, server capacity limits, and automated scaling failures have also caused major outages. Cooling failures during a London heat wave hit Google and Oracle in 2022, and lightning-induced power spikes disrupted Azure in 2018. Providers use staged rollouts, certificate tracking, traffic filtering, throttling, and backup power to mitigate risks, but no single defense covers all failure types.