Office 365 Dependent on Azure and Affected by Its Downtime
Data compiled by Gartner and Krystallize Technologies puts Azure behind Amazon Web Services (AWS) and Google in terms of uptime (availability) in North America in 2018. Although the availability of Azure is very high (99.9792%), Office 365 tenants have been affected by several recent outages that underscore the dependency Office 365 has on Azure.
Multiple Recent Azure Issues
The worst thing a cloud service can have is multiple incidents highlighted and discussed in the technology press. The January 24-25 failure of Azure Active Directory, the MFA outage in November, and the unfortunate lightening strike incident in September were followed by a DNS issue last week. Throw in the Teams outage in February caused by a problem with Azure KeyVault (the problem was with a Teams component), and there’s enough problems associated with Azure services to tarnish Azure’s image.
But let’s go back to that uptime number again. 99.9762% is extraordinarily high for a massive cloud infrastructure. No breakdown is given about which Azure components contributed to downtime, so it’s hard to know if there’s any fundamental problem with Azure or it’s just had a run of bad luck recently. I think it’s more like the latter.
Some of the problems have been down to bad luck (lightening). Others are due to a lack of resilience in the infrastructure (having a single MFA service for multiple regions), operation issues (DNS), or programming (Teams use of Azure KeyVault). In other words, there’s no single thread of weakness that causes Azure to fail. Each issue is being addressed appropriately by making services more robust and resilient, distribution across more datacenters, improvements in operational processes, or code updates.
Office 365 and Azure
When people think about Office 365 and Azure, much of the attention focuses on Azure Active Directory. Each Office 365 tenant comes with a basic Azure Active Directory subscription (to allow Office 365 to store information about accounts and settings). But the relationship between Office 365 and Azure is much wider and deeper. Some applications, like Teams, consume dozens of Azure micro-services to avoid the need to recreate wheels. Instead of building monolithic applications like Exchange and SharePoint (based largely because of their on-premises heritage), new applications seek to reuse the Office 365 software parts bin. The result is a mass of dependencies.
The availability of Azure should concern Office 365 tenants. Even with the recent glitches, Azure is still doing pretty well. And as GeekWire says “even if Microsoft lags AWS and Google in reliability scoring, unless your company is blessed with world-class operations talent, Microsoft is likely still better at operating data centers than most companies managing their own servers.”
I think you need a “but” or “however” in this paragraph opening: “The availability of Azure should concern Office 365 tenants. Even with the recent glitches, Azure is doing pretty well.” As in, “The availability of Azure should concern Office 365 tenants, but even with the recent glitches, Azure is doing pretty well.”