Loading status…
Snowflake status · Data & Backend · checked (4s ago)
Partly — Snowflake is up but running slower than usual (Partially Degraded Service).
According to Snowflake's official status page, Snowflake was reporting “Partially Degraded Service” when isuptime last checked at . Over the last 30 days, Snowflake's feed reported no outage in 100.00% of our checks.
isuptime cannot fix Snowflake, and neither can you. These are the things that are actually worth doing until it comes back.
Everything above is read from Snowflake's own status feed. This is what other visitors are saying in the last 24 hours — useful when the feed is green and you are still stuck, and never used to change the status above.
Nobody has reported Snowflake in the last 24 hours. Yours would be the first.
Reports are anonymous counts and are kept for 24 hours, then deleted. No account, no message, nothing stored about who reported. One report per service per hour, changeable until the hour is out.
Snowflake is a cloud data warehouse — storage, compute warehouses and the SQL engine over them — running on top of AWS, Azure and GCP rather than its own hardware. Analytics, BI dashboards and data pipelines sit on it, so an outage is felt as reports and jobs failing rather than a website going down.
Snowflake's status page reports per deployment: each account lives in one cloud provider and one region, and those are the components to read. An incident on AWS us-east-1 says nothing about an Azure West Europe account. Because Snowflake runs on other people's clouds, its incidents are sometimes downstream of an AWS, Azure or GCP problem.
isuptime reads this feed on a schedule and reports what it says — how that works, in detail.
One bar per day. Each shows the share of that day's checks where Snowflake's own status feed reported no outage — degraded counts as no-outage. Days with no bar are days isuptime recorded no checks, not days Snowflake was down.
13 other services isuptime tracks were also reporting a problem as of the same check, 4s ago. They are checked on one schedule, so the timing is comparable — but overlapping in time does not mean they share a cause.
Each point is one check of Snowflake's status endpoint.
Snowflake has published 50 incidents in the last 12 months, the 20 most recent shown here. Only what Snowflake posted itself — not every problem users hit.
resolved · Sep 14, 04:48 AM (6d ago)
postmortem · Sep 11, 07:18 AM (8d ago)
resolved · Sep 9, 09:53 PM (10d ago)
postmortem · Sep 4, 07:34 AM (15d ago)
postmortem · Sep 1, 09:58 AM (18d ago)
postmortem · Aug 31, 05:20 PM (19d ago)
identified · Aug 28, 12:07 AM (23d ago)
postmortem · Aug 27, 07:05 PM (23d ago)
postmortem · Aug 25, 03:53 AM (26d ago)
postmortem · Aug 24, 04:22 PM (26d ago)
postmortem · Aug 20, 01:47 PM (30d ago)
postmortem · Aug 18, 12:58 PM (32d ago)
postmortem · Aug 14, 05:29 PM (36d ago)
postmortem · Aug 14, 03:19 PM (36d ago)
postmortem · Aug 5, 02:41 PM (45d ago)
postmortem · Aug 3, 02:11 PM (47d ago)
postmortem · Aug 3, 05:53 AM (48d ago)
postmortem · Jul 30, 03:00 AM (52d ago)
postmortem · Jul 25, 09:05 AM (56d ago)
postmortem · Jul 11, 05:41 PM (70d ago)
A query that hangs is more often a suspended warehouse resuming, or queueing behind other work, than an incident — check the warehouse state and the query history. Credit exhaustion and a spending limit both stop queries dead. Role and grant changes produce permission errors that look like the data has disappeared, and it has not.
