DBS-C01 Monitoring and Troubleshooting Practice Test 12 — 15 Questions

Question 1

A company runs a self-managed Redis cluster on Amazon EC2 for caching. The cluster has one primary and two replicas, each on c5.large instances. The application experiences high latency during peak hours. CloudWatch metrics show that the primary node's CPU utilization is consistently above 80% and the network bandwidth is near the instance limit. The replicas show moderate CPU usage. The team wants to reduce latency without increasing cost significantly. Which combination of actions should the team take? (Choose two.)

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Enable Redis Cluster Mode and distribute data across multiple shards.. Enables Redis Cluster Mode to shard data across multiple nodes, reducing CPU and network load on the primary. Option E migrates to Amazon ElastiCache for Redis with Cluster Mode, which manages scaling and reduces operational overhead. Together, they address the bottleneck without significant cost increase. Option C (read from replicas) does not reduce primary write/CPU load. Option D (upgrade instance) increases cost and may still hit network limits. Option B (add replicas) increases cost and does not reduce primary load.

Answer

Add more EC2 instances as additional replicas to offload reads.

Answer

Configure the application to read from replica nodes.

Answer

Upgrade the primary to a c5.2xlarge instance type.

Question 2

A company uses Amazon DocumentDB (with MongoDB compatibility) for its content management system. The application runs on EC2 instances and connects to a DocumentDB cluster with one instance (db.r5.large). Recently, users reported that retrieving documents takes longer than usual. CloudWatch metrics show that the CPU utilization of the DocumentDB instance is at 90% and the freeable memory is below 100 MB. The team has verified that no query optimization is possible. Which action should the team take FIRST to improve performance?

Accepted Answer

Scale up the instance to db.r5.xlarge.. The DocumentDB instance is at 90% CPU and low freeable memory, indicating the instance itself is the bottleneck. Scaling up to a larger instance type (db.r5.xlarge) provides more CPU and memory, directly addressing the resource exhaustion.

Answer

Add a read replica instance to offload read traffic.

Answer

Create additional indexes on frequently queried fields.

Answer

Increase the storage volume size to improve I/O performance.

Question 3

A company runs a critical MySQL database on Amazon RDS Single-AZ (db.m5.large) with 200 GB of Provisioned IOPS (io1) storage set to 3000 IOPS. The application team reports that write operations are occasionally slow. CloudWatch metrics show that the Write IOPS metric peaks at 3500 IOPS during the slowdowns, but the average is 2000 IOPS. The Read IOPS average is 500 IOPS. The queue depth metric occasionally spikes to 20. The storage configuration includes a 50 GB General Purpose SSD (gp2) log volume attached to the same RDS instance. Which change will MOST effectively resolve the write latency?

Accepted Answer

Increase the provisioned IOPS on the io1 volume to 4000.. The io1 volume is provisioned at 3000 IOPS, but CloudWatch shows Write IOPS peaking at 3500 IOPS — exceeding the provisioned limit. When IOPS demand exceeds provisioned IOPS on io1, the volume queues I/O requests, which is confirmed by the queue depth spike to 20. Increasing provisioned IOPS to 4000 raises the ceiling above the observed peak, directly resolving the write latency.

Answer

Change the storage type to gp3 with 3000 baseline IOPS.

Answer

Move the log volume to the same io1 volume to reduce I/O overhead.

Answer

Enable Multi-AZ for failover protection.

Question 4

A company uses Amazon DynamoDB Global Tables for a multi-region application. The table is configured with on-demand capacity. The application writes data in the us-east-1 region and reads from us-west-2. Users in us-west-2 report that data written in us-east-1 is not appearing in us-west-2 within the expected replication latency of under 5 seconds. Instead, replication sometimes takes up to 30 seconds. Which two factors could be causing this increased replication latency? (Choose two.)

Accepted Answer

High network latency or packet loss between us-east-1 and us-west-2.. Option B is correct because DynamoDB Global Tables replicate changes over the AWS network between regions, so high network latency or packet loss on the path between us-east-1 and us-west-2 can directly delay the propagation of replicated writes and push latency beyond the expected sub-5-second window. Option E is correct because Global Tables use a replication stream to propagate writes; a large volume of writes in us-east-1 can create a backlog in that stream, causing replication to fall behind and take up to 30 seconds. Option A is not a cause of increased replication latency; reading before replication completes is simply the observable symptom of the delay, not a factor that slows replication itself. Option C is incorrect because read throttling in us-west-2 affects read requests, not the cross-region replication of writes. Option D is incorrect because on-demand capacity mode automatically scales write throughput and does not cause write throttling in us-east-1 in this scenario.

Answer

The application is reading from the us-west-2 table before replication completes.

Answer

The read capacity in us-west-2 is insufficient, causing read throttling.

Answer

The on-demand capacity mode is causing write throttling in us-east-1.

Question 5

A company uses Amazon DynamoDB for a high-traffic application. The application experiences throttling on a table with a provisioned read capacity of 1000 RCUs and write capacity of 500 WCUs. The table has a global secondary index (GSI) with its own provisioned capacity. A database specialist needs to identify potential causes of the throttling. (Choose two.)

Accepted Answer

The GSI's provisioned write capacity is insufficient for the write workload.. Throttling on a DynamoDB table with a GSI often stems from the GSI's capacity or key design. If the GSI's provisioned write capacity is too low for the write workload, or if its partition key has low cardinality causing hot partitions, the GSI cannot keep up, leading to throttling on the table. Both issues require adjusting the GSI's capacity or redesigning its key.

Answer

The application is using eventually consistent reads instead of strongly consistent reads.

Answer

The table's partition key is monotonically increasing, causing hot partitions.

Answer

The table's read capacity is insufficient for the read workload.

Question 6

A database administrator is troubleshooting an Amazon DynamoDB table that is experiencing throttling on read operations. The table has a provisioned read capacity of 1000 RCUs and stores items averaging 4 KB in size. The application uses eventually consistent reads. CloudWatch metrics show ConsumedReadCapacityUnits frequently hitting the provisioned limit, and ThrottledRequests for GetItem are occurring. Upon investigation, the administrator finds that some GetItem requests are retrieving items that are 12 KB in size. What is the MOST likely reason for the throttling?

Accepted Answer

The GetItem requests for 12 KB items consume 3 RCUs each (for eventually consistent reads), and the aggregate consumption exceeds the provisioned capacity.. DynamoDB charges read capacity based on item size rounded up to 4 KB. For eventually consistent reads, each 4 KB chunk costs 0.5 RCU. A 12 KB item thus costs 1.5 RCUs per read. If the application reads many such items, the aggregate consumption can exceed the provisioned 1000 RCUs, causing throttling. The solution is to either increase RCUs or optimize the application to read smaller items or use projections.

Answer

The application is using eventually consistent reads but should use strongly consistent reads to reduce RCU consumption.

Answer

The table's partition key is causing hot partitions, leading to throttling on specific partitions even though overall capacity is sufficient.

Answer

The table's provisioned read capacity is too low for the workload; increasing RCUs will resolve the throttling.

Question 7

A database specialist is investigating why an Amazon RDS for MySQL DB instance has been failing over frequently between Availability Zones. The RDS event log shows repeated 'Multi-AZ failover' events, but the DB instance metrics show normal CPU and freeable memory. The specialist suspects that the cause is at the storage layer. Which of the following is the MOST likely cause of the repeated failovers?

Accepted Answer

The primary DB instance is not receiving network heartbeats from the standby due to a misconfigured security group or network ACL.. Repeated Multi-AZ failovers with normal resource metrics suggest a network heartbeat issue. RDS relies on heartbeats between primary and standby; if these are blocked by security group or network ACL misconfigurations, failovers occur. The other options would typically show resource utilization anomalies or be one-time events.

Answer

The DB instance class is too small, causing the primary to become unresponsive during peak loads.

Answer

The DB instance's storage volume is experiencing I/O latency spikes that exceed the RDS heartbeat threshold, triggering a failover.

Answer

The standby DB instance is undergoing a software update, causing it to temporarily stop responding to heartbeats.

Question 8

A database specialist is managing an Amazon DynamoDB table that uses on-demand capacity mode. The application recently experienced a burst of traffic, and the specialist observes that some requests were throttled. Which CloudWatch metric should be monitored to identify the number of read requests that exceeded the table's provisioned throughput? (Choose the most appropriate metric.)

Accepted Answer

ReadThrottleEvents. ReadThrottleEvents is the specific CloudWatch metric that tracks the number of read requests that exceeded the table's provisioned throughput, resulting in throttling. In on-demand mode, DynamoDB automatically scales, but sudden bursts can still cause throttling if the request rate doubles within 30 minutes or exceeds twice the previous peak. Monitoring this metric helps the specialist identify and address read throttling issues.

Answer

ThrottledRequests

Answer

ConsumedReadCapacityUnits

Answer

ProvisionedReadCapacityUnits

Question 9

A database specialist is investigating an Amazon Aurora MySQL DB cluster where the application reports occasional spikes in read latency. The cluster has one writer and two readers. The specialist suspects that the reader instances are not distributing the read load evenly and wants to verify the current configuration and identify the root cause. Which two actions should the specialist take to diagnose the issue? (Choose two.)

Accepted Answer

Examine the Amazon CloudWatch metrics for each reader instance, specifically focusing on the DatabaseConnections and SelectLatency metrics.. To diagnose uneven read distribution in Aurora MySQL, the specialist must verify that the application connects via the reader endpoint, which balances connections, and then inspect per-reader CloudWatch metrics to see actual load. These two steps directly address whether the issue is misrouting or instance-specific performance.

Answer

Use AWS CloudTrail to inspect API calls made to the Aurora cluster and identify any recent modifications to the cluster's read replicas.

Answer

Enable and analyze the slow query log on the writer instance to identify long-running read queries that may be offloaded to readers.

Answer

Review the Aurora cluster's parameter group to ensure that the query_cache_size parameter is set appropriately for the workload.

Question 10

A database specialist is responsible for an Amazon RDS for SQL Server DB instance. The instance is configured with Multi-AZ. During a failover test, the specialist notices that the DNS endpoint took longer than expected to update, causing application downtime. Which of the following should the specialist recommend to minimize downtime during future failovers?

Accepted Answer

Enable the RDS Proxy for the DB instance to manage connections during failover.. Amazon RDS Proxy is designed to improve availability by pooling and sharing connections, and it automatically directs traffic to the new primary during a failover. It reduces the impact of DNS propagation delays because the application connects to the proxy endpoint, which remains stable. This minimizes downtime and is a best practice for Multi-AZ deployments.

Answer

Configure the application to use the RDS instance's IP address directly.

Answer

Use Amazon Route 53 with a health check and a failover routing policy to the standby instance.

Answer

Implement a custom DNS solution with a shorter TTL.

Question 11

A database specialist needs to review the slowest SQL statements executed against an Amazon Aurora MySQL DB cluster over the past week. The cluster has Performance Insights enabled with a retention period of 7 days. Which action should the specialist take to identify the top SQL statements by average latency?

Accepted Answer

Query the slow query log in Amazon CloudWatch Logs, filter for the past 7 days, and sort by Query_time descending.. The slow query log captures statements that exceed long_query_time and includes execution time. Publishing it to CloudWatch Logs allows filtering by time range and sorting by Query_time to find the slowest statements. Performance Insights shows load, not latency ranking. CloudTrail tracks API activity, and the general query log lacks timing details, so neither can rank by latency.

Answer

Open the Amazon RDS console, navigate to Performance Insights for the DB cluster, set the time range to the past 7 days, and sort the Top SQL table by Average Active Sessions.

Answer

Enable the Aurora MySQL general query log, wait 7 days, and then download the log from the RDS console to analyze statement durations.

Answer

Use AWS CloudTrail to identify the slowest SQL statements by filtering on the RDS Data API events for the cluster.

Question 12

A database specialist is troubleshooting an Amazon Aurora MySQL DB cluster that is experiencing increased write latency. The specialist suspects that the issue is related to the Aurora storage subsystem or replication. Which two CloudWatch metrics should be examined to diagnose the write latency? (Choose two.)

Accepted Answer

CommitLatency. CommitLatency and DiskQueueDepth are the two metrics that directly relate to write latency on an Aurora MySQL writer instance. CommitLatency measures the time to commit a transaction, which includes writing to the Aurora storage volume. DiskQueueDepth indicates storage I/O contention. Together, they help determine if the write latency is caused by storage bottlenecks or slow commits.

Answer

BufferCacheHitRatio

Answer

SelectLatency

Answer

AuroraReplicaLag

Question 13

A database specialist is managing an Amazon DocumentDB (with MongoDB compatibility) cluster. The application team reports that some queries are running slowly. The specialist checks the cluster's CloudWatch metrics and sees that the 'BufferCacheHitRatio' is below 90% and 'DiskQueueDepth' is elevated. The cluster has three instances: one primary and two replicas. The application reads from the primary. Which action should the specialist take to improve read performance with the LEAST operational overhead?

Accepted Answer

Increase the instance size of the primary instance to provide more memory for the buffer cache.. The low buffer cache hit ratio and high disk queue depth indicate that the instance lacks sufficient memory to cache frequently accessed data, causing excessive disk I/O. Vertically scaling the primary instance by choosing a larger instance class increases available RAM, allowing a larger buffer cache and reducing disk reads. This requires a brief failover but no application changes, making it the least operational overhead solution.

Answer

Add a read replica and direct read queries to it to distribute the load.

Answer

Enable the 'IndexOnly' query planner option to force queries to use indexes.

Answer

Modify the cluster to use a larger volume size to increase IOPS.

Question 14

A database specialist is troubleshooting an Amazon Aurora PostgreSQL cluster that is experiencing increased read latency on its reader instances. The cluster has one writer and two readers. The specialist has already confirmed that the writer instance is not overloaded and that the readers are not experiencing high CPU. Which two actions should the specialist take to diagnose the issue? (Choose two.)

Accepted Answer

Check the Amazon CloudWatch metric 'AuroraReplicaLag' to see if the readers are lagging behind the writer.. Increased read latency on Aurora readers can be caused by replication lag or inefficient queries. Checking AuroraReplicaLag helps determine if readers are behind, and reviewing pg_stat_activity identifies long-running queries that may be causing contention. These two actions directly address common causes.

Answer

Verify that the reader instances are in the same Availability Zone as the writer to reduce network latency.

Answer

Check the 'FreeableMemory' metric for the reader instances to see if they are running low on memory.

Answer

Examine the 'VolumeReadIOPs' metric for the reader instances to identify if there is an I/O bottleneck.

Question 15

A company uses Amazon DynamoDB with a table that has a partition key of CustomerID and a sort key of OrderDate. The table is provisioned with 1000 RCUs and 1000 WCUs. During a flash sale, many requests for a single customer fail with ProvisionedThroughputExceededException, while other customers' requests succeed. What is the most likely cause?

Accepted Answer

The requests for that customer are exceeding the per-partition throughput limit.. DynamoDB partitions data by partition key value. All items with the same CustomerID reside in the same partition, which has a maximum write throughput of 1000 WCUs. During the flash sale, one customer's write activity exceeds that limit, causing throttling for that partition only. Other customers are unaffected because their data is in different partitions. The solution is to distribute the workload more evenly, such as by adding a random suffix to the partition key.

Answer

The table's sort key is causing uneven data distribution.

Answer

The application is using eventually consistent reads instead of strongly consistent reads.

Answer

The table's provisioned throughput is too low for the overall traffic.