Snowflake update caused a blizzard of failures worldwide

Customers in 10 of the company’s 23 regions had “operations fail or take an extended amount of time to complete.”

Snowflake pushed an update this week that caused a “major outage” worldwide, leaving many users unable to query data, experiencing failures when ingesting files, and receiving error messages for 13 hours, the company wrote in an impact statement.

“Our initial investigation has identified that our most recent release introduced a backwards-incompatible database schema update,” the company wrote in its incident report. “As a result, previous release packages errantly referenced the updated fields, resulting in version mismatch errors and causing operations to fail or take an extended amount of time to complete.”

The update took down 10 of Snowflake’s 23 global regions. In the US, the outage hit chunks of Snowflake customers accessing its Azure data center in Virginia, and its AWS data center in Oregon. Globally, users of Snowflake’s data centers in Ireland, Zurich, London, Sweden, Mumbai, Singapore, and Mexico were hit.

The outage was first reported as an issue with Snowflake Data Cloud at 0255 UTC Tuesday when customer queries against their data returned SQL execution errors.

About an hour and a half later, Snowflake said that it had identified the problem and it was working on a fix. The systems were back up and running by about 0500 UTC.

“I dealt with this yesterday,” one user groused on Snowflake's Reddit page. “Took them way too long to role [sic] back this change”

The company said it would release its root cause analysis within five days of closing the incident, which should be Sunday.

“We have nothing additional to share at this time,” a spokesperson told The Register.

This was Snowflake’s second incident inside of a week after a Dec 10 database infrastructure issue caused degraded performance among users who employ its software out of the AWS datacenter in Oregon, according to its status page.

Snowflake is not the only data platform with issues this month.

Between Dec 11 and Dec 13, Snowflake’s archrival Databricks had a days-long service disruption where customers in several US regions experienced latency or errors when using Mosaic AI, the company’s assistant for building AI agents, which was contained to users of the company’s Azure data centers in seven US regions.

That came after customers in its US Gov West AWS region endured a “complete outage, with all Databrick services presumed unavailable” for about two hours on Dec 3.

Unlike Snowflake, which lists the preliminary cause of outages and follows up with a root cause analysts on its incident page, Databricks does not publish the cause of its outages.

Databricks did not respond to an email from The Register seeking comment. ®