Snowflake for SQL Server Users - Part 19 - Fail-Safe

In the previous post in this series, I talked about how historical data in Snowflake can be accessed by a concept called Time Travel. But Time Travel isn’t the only way to get to historical data.

In Snowflake, Fail-safe is a system that provides protection against system failures and other nasty events like hardware failures or serious security breaches.

Immediately after the Time Travel retention period ends, a Fail-safe period begins.  It lasts for a non-configurable period of seven days.

Fail-safe isn’t intended for users to access like they do with Time Travel. Instead, Fail-safe data is held so that Snowflake support can recover data that has been lost or damaged.

What about backup?

With Snowflake, there is no concept of a standard backup. The argument is that their multi-data center and highly redundant architecture almost removes the need for backup. And the thinking is that Fail-safe removes that final risk.

The only issue I see with that logic, is the same as with most cloud-based databases and data warehouses: Users sometimes want to keep point in time backups over long periods. I don’t currently see any option available for this within Snowflake. But as I mentioned, they aren’t alone in that.

For an index to all posts in this series, see the first post here.

2019-12-13