Data Warehousing on AWS Training Data Warehousing on AWS: Workload Management 4 — Questions and Answers
Question 1: How does enabling 'Enable short query acceleration' affect the WLM default queue in Redshift?
- It merges the default queue with the first user-defined queue
- SQA creates a separate fast lane outside of standard WLM queues for eligible short queries (Correct answer)
- It reduces the default queue's memory allocation by 50%
- It disables the default queue entirely
Correct answer: SQA creates a separate fast lane outside of standard WLM queues for eligible short queries
SQA operates independently of configured WLM queues by routing predicted short-running queries to a dedicated acceleration lane, leaving standard queues for longer workloads.
Question 2: Which parameter group setting must be modified to apply WLM configuration changes to an Amazon Redshift cluster?
- Cluster security group
- Cluster parameter group (Correct answer)
- Cluster subnet group
- IAM instance profile
Correct answer: Cluster parameter group
WLM configuration is stored in the cluster parameter group; changes to the parameter group are applied to associated clusters (requiring a reboot for static parameters).
Question 3: What is the recommended approach when you need a dedicated queue for ETL jobs that run during off-peak hours in Redshift?
- Create a separate Redshift cluster exclusively for ETL
- Configure a WLM queue with user group routing mapped to the ETL database users (Correct answer)
- Use Amazon EMR instead of Redshift for ETL workloads
- Schedule ETL queries using Amazon EventBridge with no WLM changes
Correct answer: Configure a WLM queue with user group routing mapped to the ETL database users
Creating a dedicated WLM queue with user group routing for ETL database users ensures ETL jobs are isolated from interactive query workloads in the same cluster.
Question 4: Which Redshift system view shows currently queued and executing queries along with their WLM state?
- STV_WLM_SERVICE_CLASS_STATE (Correct answer)
- STV_RECENTS
- SVL_QLOG
- STL_QUERY
Correct answer: STV_WLM_SERVICE_CLASS_STATE
STV_WLM_SERVICE_CLASS_STATE provides a real-time snapshot of each WLM service class (queue) including current concurrency, queued queries, and resource usage.
Question 5: In Redshift WLM, what is a 'service class'?
- An IAM role that grants access to specific Redshift tables
- The internal representation of a WLM queue including its resource allocations (Correct answer)
- A tiered pricing model for Redshift Reserved Nodes
- A CloudWatch alarm category for Redshift performance metrics
Correct answer: The internal representation of a WLM queue including its resource allocations
In Redshift internals, each WLM queue is referred to as a service class, identified by a service class ID that appears in system tables like STV_WLM_SERVICE_CLASS_STATE.
Question 6: When applying a new WLM configuration that changes only dynamic properties in Redshift, what is required?
- A full cluster reboot
- No reboot is needed; changes take effect immediately (Correct answer)
- A manual vacuum and analyze operation
- Terminating all active sessions before applying
Correct answer: No reboot is needed; changes take effect immediately
Dynamic WLM properties such as concurrency level and memory allocation can be applied without rebooting the cluster, taking effect for new queries immediately.
Question 7: What is the default behavior for queries submitted by superusers in Redshift WLM queue routing?
- Superuser queries are always routed to the queue with the most free slots
- Superuser queries are automatically routed to the dedicated superuser queue (Correct answer)
- Superuser queries bypass WLM entirely and run at full cluster capacity
- Superuser queries are subject to the same routing rules as regular users
Correct answer: Superuser queries are automatically routed to the dedicated superuser queue
Redshift maintains a reserved superuser queue that superuser queries are automatically routed to, ensuring administrative queries are never blocked by regular workload queues.
How does enabling 'Enable short query acceleration' affect the WLM default queue in Redshift?