For about three hours on Thursday, Apple had a widespread cloud outage that affected services including the App Store, Apple Music, and Apple TV.
Multi-tenant cloud is supposedly more resilient than in-house infrastructure.
But events in the past month have reminded a world that’s moved wholesale to the cloud that sometimes that resiliency fails. And that minutes and hours count, for example, when you need to authenticate to vital online services.
Apple’s outage follows two catastrophic failures in the past month at CDN provider Cloudflare, an outage and disruptions affecting Google Cloud customers, and availability problems across Facebook, WhatsApp and Instagram just days ago.
Apple’s system status page indicates it had about a three-hour outage yesterday, which affected the App Store, Apple Books, Apple ID, Apple Music, Apple Music Subscriptions, Apple Online Store, Apple Pay, Apple TV, the Apple device enrollment service, and more. Users couldn’t even use Find My iPhone to locate their devices.
An outage this widespread across Apple services was reportedly unprecedented and nationwide, according to one user’s account of an interaction with Apple store staff.
ThousandEyes, a US-based network monitoring startup, had a good explanation of Cloudflare’s outage that was caused by a Verizon BGP internet-routing blunder. It says Apple’s cloud collapse was also due to a ‘BGP route flap’ issue that caused severe packet loss for users in North America.
“The packet loss appears to have been caused by a BGP route flap issue, where a routing announcement is made and withdrawn in quick succession, often repeatedly,” ThousandEyes wrote.
The extent of Apple’s outage gave iPhone users a brief reminder of how dependent they are on Apple’s cloud, which affected their ability to pay at the checkout, buy stuff online, sign in to apps, use their paid-for music subscriptions, and carry out work.
The recent spate of cloud outages has also shown different companies approaches to explaining to customers about what went wrong. Cloudflare and Google, which provide services to businesses, offered detailed technical accounts.
Facebook this week apologized for its issues but only explained that a “routine maintenance operation” triggered an issue that prevented people from sending photos and videos.
Microsoft has had its share of cloud outages too this year, preventing businesses across the world from using Azure, Microsoft 365, Dynamics, and DevOps, and previously blocking users from authenticating to its core cloud services.
ZDNet has contacted Apple for an explanation of what happened yesterday.