Azure SQL Database caused immoderate annoyance complete nan play pinch admins connected nan US eastbound seashore incapable to link to nan work pursuing a web infrastructure powerfulness failure.
The disruption started astatine 07:33 UTC connected September 16 and wasn't afloat mitigated until much than 12 hours later astatine 21:38, Microsoft confirmed connected its status page.
Microsoft said: "Some customers utilizing Azure SQL Database successful East US whitethorn person knowledgeable issued erstwhile trying to link to Databases. Connections to database hosted successful East US region whitethorn person resulted successful persistent errors aliases timeouts."
As a customer noted connected nan level formerly known arsenic Twitter: "Sql DB relationship mislaid successful production. The failover doesn't footwear in, causing effect successful regular activities. Thanks deity it (sic) Saturday."
The nonstop origin of nan wobbly sessions has yet to beryllium determined. Microsoft said: "We identified that during a little play of clip an underlying web infrastructure knowledgeable powerfulness power issue. This caused compute nodes to go unhealthy, resulting successful failures and timeouts for SQL Database."
- Google exec: Microsoft Teams concession 'too little, excessively late'
- Cloud is present to stay, but customers are starting to mobility nan cost
- Microsoft 'fesses to codification blunder successful Azure Container Apps
- Microsoft whips up unrest aft revealing Azure AD sanction change
As is often nan case, nan domino effect kicked in, and downstream services that trust connected SQL Database struggled to tally arsenic normal.
"We were notified of this rumor done our soul monitoring systems, prompting america to initiate a thorough investigation," Microsoft said. "In bid to mitigate nan first impact, we rebooted nan affected compute nodes, thereby restoring functionality to mostly of databases. Subsequently, nan remaining SQL DB instances was brought backmost online sewage bring survives to afloat functionality."
The powerfulness rumor astatine nan guidelines of nan degraded services is still nan root of investigations by engineers astatine Microsoft arsenic they effort to "establish a workstream to forestall early occurrence."
So not precisely nan mother of each outages that caused customers to gnash their teeth aliases propulsion their hairsbreadth out, but besides not a awesome look for Microsoft SQL DBV connected Azure. Still, it wasn't arsenic embarrassing arsenic nan incident reported a week ago that wiped retired Azure services connected nan Australian East unreality region.
In this, "a inferior powerfulness sag tripped a subset of nan cooling units offline successful 1 datacenter, wrong 1 of nan Availability Zones." Microsoft concluded that understaffing and automation led to nan challenge. ®