This week, Snowflake experienced a significant outage following a software update, resulting in a global service disruption. Users encountered issues with data queries and file ingestion, experiencing errors for approximately 13 hours. The company’s incident report indicated that a newly released database schema update was incompatible with earlier versions, causing operational failures.
Impact of the Outage
The update impacted 10 out of Snowflake’s 23 global regions. Notable areas affected include:
- Azure data center in Virginia, USA
- AWS data center in Oregon, USA
- Data centers in Ireland
- Zurich
- London
- Sweden
- Mumbai
- Singapore
- Mexico
The outage was first reported at 02:55 UTC on Tuesday, when users encountered SQL execution errors. Snowflake identified the issue approximately 90 minutes later and began working on a resolution. Services were restored around 05:00 UTC.
User Reactions
User frustration was evident, with comments on Snowflake’s Reddit page expressing dissatisfaction with the speed of the resolution. One user remarked on the lengthy rollback process required to address the issue.
Future Insights
Snowflake has committed to releasing a root cause analysis within five days of the incident’s conclusion. A spokesperson indicated that more information would be available later this week.
This incident follows another issue reported on December 10, which led to degraded performance for users connected to the AWS data center in Oregon.
Industry Context
Snowflake is not alone in facing challenges this month. Its rival, Databricks, encountered its own significant service interruptions. From December 11 to 13, several US users faced latency and errors when using the Mosaic AI tool. This disruption followed a total outage in the US Gov West AWS region on December 3.
Unlike Snowflake, Databricks does not publicly disclose the causes of its outages, leaving users without detailed explanations of service disruptions.
