ADA-C01 Question 554
Select 2Implement best practices for backups in SnowflakeA financial services company stores critical transaction data in Snowflake and must be able to recover from accidental table drops, user errors that go unnoticed for several weeks, and a rare regional outage. The company wants to follow Snowflake backup best practices while minimizing operational overhead. Which TWO actions should the Snowflake administrator recommend?
- A
Enable an appropriate DATA_RETENTION_TIME_IN_DAYS value on critical permanent objects so Time Travel can support short-term recovery needs such as accidental deletes or updates.
- B
Create replication or failover groups for the business-critical databases and account objects in a secondary region/cloud so the company has a disaster recovery copy outside the primary region.
- C
Convert critical permanent tables to transient tables because transient objects reduce storage cost and still provide Fail-safe for long-term recovery.
- D
Schedule a nightly COPY INTO to an internal stage and treat the staged files as the primary long-term backup strategy for Snowflake metadata and account objects.
- E
Rely on Snowflake Fail-safe as the main business continuity strategy because it is designed for customer-initiated operational recovery and rapid self-service restores.
Show answer and explanation
Correct answers: A, B
Explanation
Snowflake backup best practices are layered rather than based on a single feature. For short-term logical recovery from mistakes such as accidental updates, deletes, or dropped tables, use Time Travel with an appropriate DATA_RETENTION_TIME_IN_DAYS setting on critical permanent objects. For broader disaster recovery, especially protection from regional outages, use database replication and failover groups to maintain a recoverable copy in another region or cloud provider where supported. Fail-safe exists, but it is a last-resort recovery mechanism managed through Snowflake and should not be treated as the main backup or continuity strategy. Also, transient objects do not have Fail-safe, so using them for critical data weakens recoverability. While unloading data with COPY INTO may support external archival requirements, it does not replace Snowflake-native recovery capabilities for full operational backup and DR planning. These recommendations align with Snowflake documentation on Time Travel, Fail-safe, replication, and failover groups.
- A. Correct.
Correct. Time Travel is a core Snowflake recovery capability and is a best-practice component of short-term backup/recovery planning. Setting DATA_RETENTION_TIME_IN_DAYS appropriately on critical permanent objects helps recover from accidental DML/DDL changes within the configured retention period. This addresses user errors such as mistaken updates, deletes, or dropped objects detected within that window.
- B. Correct.
Correct. Replication/failover groups are the recommended Snowflake-native approach for disaster recovery across regions or clouds. They provide a secondary copy of supported databases and certain account-level objects, which is important for rare but high-impact regional outages. This aligns with best practices for business continuity beyond local logical recovery.
- C. Incorrect.
Incorrect. Transient tables do not provide Fail-safe. In Snowflake, transient and temporary objects are designed to reduce storage costs partly by not having Fail-safe protection. Converting critical permanent tables to transient would weaken recoverability rather than improve it.
- D. Incorrect.
Incorrect. Exporting data with COPY INTO can be useful for specific archival or interoperability use cases, but it is not a complete backup strategy for Snowflake. It does not preserve full Snowflake object state, history, governance configuration, or account metadata the way native recovery, cloning, and replication features address them. Treating staged files as the primary backup strategy is a common misconception.
- E. Incorrect.
Incorrect. Fail-safe is not intended as a customer-managed, operational recovery mechanism. It is a last-resort Snowflake recovery feature after Time Travel expires and generally requires Snowflake Support involvement. It is not designed for rapid self-service business continuity or as the primary recovery plan.