Cloud Storage and Dataflow: How to Achieve a 15-Minute RPO with Pub/Sub?
You have designed an Apache Beam processing pipeline that reads from a Pub/Sub topic. The topic has a message retention duration of one day, and writes to a Cloud Storage bucket. You need to select a bucket location and processing strategy to prevent data loss in case of a regional outage with an RPO of 15 minutes. What should you do?
Community Votes
82% of anonymous learners picked answer D. Votes are pick records left by other test-takers — they are not the verified answer.
Community Insight
This question tests whether you know that default dual-region replication is asynchronous and not sufficiently fast for a 15-minute RPO; the correct answer flags turbo replication as the feature that guarantees this RPO.
Learn how to configure a Cloud Storage bucket and Apache Beam pipeline to prevent data loss during a regional outage. Community consensus is that a dual-region bucket with turbo replication is required to meet a 15-minute RPO, combined with Pub/Sub seek-replay.
Selecting a plain dual-region bucket (option A) because dual-region sounds geo-redundant; however, without turbo replication, Cloud Storage can take hours or up to 12 hours to replicate data, failing the RPO requirement.
Community Discussion (9 comments)
- Low RPO: Dual-region buckets offer synchronous replication, ensuring data is immediately available in both regions, aligning with the 15-minute RPO. - Turbo Replication: enabling turbo replication can further reduce replication latency to near-real-time for even stricter RPO requirements. - Resilient Data Storage: Dual-region buckets ensure data availability even during regional outages, protecting processed data. - Fast Recovery: Reprocessing from the last 15 minutes of acknowledged messages minimizes data loss and downtime.
Comments & Corrections
No comments yet — spotted an error or have a note? Share it below.
Expert Analysis
Why the Answer Is Correct
Option D is correct because a dual-region Cloud Storage bucket with turbo replication is explicitly designed to provide an RPO of 15 minutes for 100% of newly written objects. As one commenter noted, the official docs state that turbo replication ensures replication to both regions within 15 minutes regardless of object size.
A dual-region bucket without turbo replication (A) does not guarantee this RPO; its replication is eventual and can take up to 12 hours in the worst case. Therefore, only D satisfies the stated 15-minute RPO.
Additionally, the full strategy includes seeking the Pub/Sub subscription back in time and restarting the Dataflow job in a secondary region, which is essential for recovering acknowledged messages that may not have been fully written to the bucket yet.
Why the Other Options Are Wrong
Option A is wrong because it relies on a plain dual-region bucket without turbo replication, which cannot meet a 15-minute RPO. The replication lag could be hours, causing data loss in an outage.
Option B (multi-regional) may provide geo-redundancy but is not the specific recommended strategy for meeting a low RPO with Dataflow and Pub/Sub replay; it also lacks the turbo replication guarantee.
Option C (regional) stores data in a single region, so a regional outage would make the output unavailable and likely cause permanent data loss.
Community comments supporting A often mistake default dual-region replication for synchronous replication, but official documentation shows turbo replication is needed for the 15-minute RPO.
Community Comment Notes
Comment [1] (likes=8) outlines the full winning workflow: enable turbo replication, monitor with Cloud Monitoring, seek the subscription back 60 minutes, and start in a secondary region. This matches option D.
Comment [2] (likes=5) correctly explains that a plain dual-region bucket without turbo replication takes at least 1 hour to sync and has a 12-hour SLA for full replication, so it fails the RPO.
Comment [3] (likes=4) directly quotes the Cloud Storage documentation: turbo replication is designed to replicate 100% of newly written objects within 15 minutes.
Comments [5] and [6] argue that turbo makes the seek step unnecessary, but they ignore that seek-replay covers in-flight Dataflow data, not just the bucket replication; both mechanisms work together.
Official Reference
Exam Strategy
When you see "RPO of 15 minutes" in a Dataflow/Cloud Storage scenario, immediately look for the phrase "turbo replication" in the answer options. Also remember that a complete disaster recovery setup always includes a Pub/Sub seek-back step to replay acknowledged messages, so don't choose an answer that only discusses the bucket type.