Courseiva

AWS Certified Database Specialty DBS-C01 (DBS-C01) — Questions 451525

1663 questions total · 23pages · All types, answers revealed

Page 6

Page 7 of 23

Page 8
451
MCQmedium

A company runs a MongoDB-compatible workload on Amazon DocumentDB. They notice that many read requests are returning stale data even though reads are directed to the primary instance. What is the MOST likely cause?

A.The application's session is pinned to a secondary replica despite requesting the primary.
B.The application is using a read preference that allows secondary reads.
C.The primary instance is experiencing high CPU utilization, causing delayed writes.
D.The storage volume is using the default eventually consistent configuration for primary reads.
AnswerB

If the read preference is set to 'secondaryPreferred' or similar, reads may go to secondary replicas which are eventually consistent.

Why this answer

The most likely cause of stale reads from the primary instance is that the application is using a read preference that allows secondary reads. In Amazon DocumentDB, even if the connection string specifies the primary endpoint, the MongoDB driver's read preference setting (e.g., `secondaryPreferred` or `nearest`) can cause reads to be served from replica instances, which may have replication lag and thus return stale data. The default read preference is `primary`, but if the application explicitly sets a different preference, reads can be directed to secondaries without the developer realizing it.

Exam trap

The trap here is that candidates assume connecting to the primary endpoint always guarantees primary reads, but the MongoDB driver's read preference setting can silently redirect reads to secondaries, causing stale data even when the endpoint is correct.

How to eliminate wrong answers

Option A is wrong because session pinning to a secondary replica does not occur when the application explicitly requests the primary; DocumentDB's replica set driver handles failover and read preference, not arbitrary pinning. Option C is wrong because high CPU utilization on the primary delays writes but does not cause stale reads on the primary itself; stale reads occur only when reading from a secondary with replication lag. Option D is wrong because DocumentDB uses a single, strongly consistent storage volume for all instances in the cluster; there is no 'eventually consistent configuration' for primary reads, and primary reads are always strongly consistent.

452
MCQmedium

A company uses Amazon Redshift for data warehousing. They run a daily ETL job that loads data into the cluster. Recently, the job started failing with 'Disk Full' errors. The cluster has 5 RA3 nodes. Which step should be taken to resolve the issue?

A.Disable concurrency scaling to free up resources
B.Run a VACUUM command to reclaim space from deleted rows
C.Resize the cluster to a larger node type or add more nodes
D.Enable Redshift Spectrum to offload queries to S3
AnswerC

RA3 nodes separate compute and storage; you can increase storage by resizing or adding nodes.

Why this answer

Resize the cluster to a larger node type or add more nodes. RA3 nodes use managed storage with a local cache, and a 'Disk Full' error typically indicates that the local cache or the overall storage limit for the cluster is exhausted. Resizing adds more local cache and increases the total managed storage capacity, resolving the disk full error.

Option A is incorrect because disabling concurrency scaling does not free up disk space; it only affects query concurrency. Option B is incorrect because VACUUM reclaims space from deleted rows, but it requires temporary disk space; if the disk is already full, VACUUM may fail or not help. Option D is incorrect because Redshift Spectrum allows querying data directly from S3 without loading it into Redshift, but it does not resolve a disk full error on the Redshift cluster itself.

453
Multi-Selecthard

Which TWO settings should be verified when troubleshooting an RDS for MySQL instance that has a high number of aborted connections? (Choose 2.)

Select 2 answers
A.connect_timeout parameter
B.max_allowed_packet parameter
C.query_cache_type parameter
D.binlog_retention_hours parameter
E.max_connections parameter
AnswersA, B

Low connect_timeout can cause aborted connections if client takes too long.

Why this answer

Options A (connect_timeout) and B (max_allowed_packet) are correct because a low connect_timeout can cause connections to abort if not completed within the time limit, and a low max_allowed_packet can cause large queries to fail, resulting in aborted connections. Option C is incorrect because query_cache_type affects caching, not connection handling. Option D is incorrect because binlog_retention_hours is for binary log retention, unrelated to connections.

Option E is incorrect because max_connections limits the total number of simultaneous connections but does not directly cause aborted connections; excessive connection attempts might be due to other issues.

454
MCQhard

Based on the CLI output, what is true about this RDS instance?

A.The instance runs Amazon Aurora PostgreSQL
B.The instance is a Multi-AZ deployment
C.The instance is a Read Replica of another RDS instance
D.The instance uses Provisioned IOPS (io1) storage
AnswerC

ReadReplicaSourceDBInstanceIdentifier is set.

Why this answer

The CLI output shows `ReplicaLag` with a value of `0`, which is a field that only appears when the RDS instance is configured as a Read Replica. A Read Replica maintains asynchronous replication from a source DB instance, and the lag metric indicates how far behind the replica is. Since the output includes this field, the instance must be a Read Replica.

Exam trap

The trap here is that candidates see `ReplicaLag: 0` and assume it means no replication is happening or that it indicates a Multi-AZ setup, but in reality, a lag of 0 simply means the replica is fully caught up, and the presence of the field itself confirms it is a Read Replica, not a Multi-AZ standby.

How to eliminate wrong answers

Option A is wrong because the output does not show any Aurora-specific fields (e.g., `DBClusterIdentifier`, `AuroraReplicaLag`) and the engine would be listed as `aurora` or `aurora-postgresql`, not a standard RDS engine. Option B is wrong because a Multi-AZ deployment does not expose a `ReplicaLag` field; Multi-AZ uses synchronous replication and the replica is not directly accessible for reads. Option D is wrong because the output does not include `StorageType` set to `io1` or `ProvisionedIOPS`; without those fields, we cannot conclude the instance uses Provisioned IOPS storage.

455
MCQhard

A financial services company uses Amazon DynamoDB to store sensitive customer data. The security team requires that all data at rest be encrypted using a customer-managed AWS KMS key (CMK) with automatic rotation enabled. The DynamoDB table was created with the default AWS-managed key. Which steps are necessary to transition to a customer-managed CMK while minimizing downtime and data loss?

A.Modify the DynamoDB table to update the encryption key to the new CMK using the AWS Console.
B.Export the table data to Amazon S3, create a new DynamoDB table with the new CMK, import the data, and update the application to use the new table.
C.Disable encryption at rest, then re-enable it with the new CMK.
D.Update the KMS key policy to grant DynamoDB access to the new CMK, then rotate the key.
AnswerB

This is the only way to change the encryption key, as DynamoDB does not allow in-place key changes.

Why this answer

DynamoDB does not support in-place modification of the encryption key for an existing table. To transition from an AWS-managed key to a customer-managed CMK, you must export the table data to Amazon S3, create a new DynamoDB table configured with the new CMK, import the data, and update the application to point to the new table. This approach minimizes downtime by allowing the original table to serve reads/writes during the export and import process, and avoids data loss by using DynamoDB's native export and import features.

Exam trap

The trap here is that candidates assume DynamoDB allows in-place encryption key changes (like some other AWS services), but DynamoDB requires a table recreation to change the encryption key, making the export/import workflow necessary.

How to eliminate wrong answers

Option A is wrong because the AWS Console does not allow modifying the encryption key of an existing DynamoDB table; encryption settings can only be set at table creation time. Option C is wrong because DynamoDB does not support disabling encryption at rest on an existing table; encryption is always enabled and cannot be toggled off or changed in place. Option D is wrong because updating the KMS key policy or rotating the key does not change the encryption key used by the table; the table continues to use the originally assigned key, and key rotation only affects future encryption operations, not the key used for existing data.

456
MCQhard

A company is migrating a 3 TB on-premises Oracle database to Amazon Aurora PostgreSQL. The source database runs Oracle 12c on a Linux server with a 1 Gbps network connection to AWS via Direct Connect. The migration must have minimal downtime and be completed within a 2-day window. The database is heavily used during business hours (9 AM - 5 PM) and has low activity overnight. The company has a test environment on AWS already. The migration team plans to use AWS DMS with CDC for ongoing replication. They also plan to use AWS SCT for schema conversion. They start the migration on a Friday evening. On Saturday morning, the CDC replication lag is increasing, and the target Aurora instance is struggling to keep up. The team notices that the source database has a high number of write transactions even during the weekend, and the DMS replication instance is a small instance type. Which action should the team take to complete the migration on time?

A.Reduce the number of tables being migrated to lower the load.
B.Upgrade the DMS replication instance to a larger size to increase throughput.
C.Pause the migration and restart it on Monday during low activity.
D.Switch to use AWS Snowball to transfer data and bypass DMS.
AnswerB

Larger instance can handle more write transactions.

Why this answer

Upgrading the DMS replication instance increases its capacity to handle high write volume. Option A is wrong because it doesn't address the bottleneck. Option C is wrong because it adds complexity and may not help.

Option D is wrong because it doesn't solve the replication lag.

457
MCQmedium

A database administrator needs to audit all SQL statements executed on an Amazon RDS for Oracle DB instance. The audit logs must be stored in Amazon S3 for long-term retention and analysis with Amazon Athena. Which solution meets these requirements?

A.Enable Enhanced Monitoring on the RDS instance and publish logs to CloudWatch Logs.
B.Configure Oracle's unified auditing and stream audit logs to CloudWatch Logs, then export to S3.
C.Enable AWS CloudTrail to capture RDS API calls and store them in S3.
D.Enable detailed billing reports and configure them to include database queries.
AnswerB

Oracle's unified auditing can be configured to stream audit logs to CloudWatch Logs, which can then be exported to Amazon S3 for long-term retention and analysis with Athena.

Why this answer

Oracle's unified auditing can be configured to stream audit logs to CloudWatch Logs, which can then be exported to Amazon S3 for long-term retention and analysis with Athena. Option A is incorrect because Enhanced Monitoring provides OS-level metrics, not SQL audit logs. Option C is incorrect because AWS CloudTrail captures API calls to RDS, not SQL statements executed within the database.

Option D is incorrect because detailed billing reports do not include database query logs.

458
MCQmedium

Refer to the exhibit. An IAM policy is attached to a user. What is the effect when the user attempts to delete the RDS DB instance named 'prod-db'?

A.The user can delete any other instance except 'prod-db'.
B.The user can delete the instance because the Deny statement only applies to snapshots.
C.The user cannot delete the instance because of the explicit Deny statement.
D.The user can delete the instance because of the Allow on DescribeDBInstances.
AnswerC

Explicit Deny overrides Allow.

Why this answer

The policy explicitly denies the rds:DeleteDBInstance action on the specific resource. Even though there is an Allow on other actions, an explicit Deny overrides any Allow. The user cannot delete the instance.

Option A is wrong because the Deny takes precedence. Option B is wrong because the policy explicitly prevents deletion. Option D is wrong because the Deny is on the specific instance.

459
MCQhard

A company is deploying a multi-region application with Amazon Aurora Global Database. They need to ensure that the secondary region can be promoted to primary with minimal data loss in the event of a regional failure. Which configuration should they use?

A.Deploy Amazon Aurora Serverless with cross-Region replication.
B.Deploy Amazon Aurora Global Database with one primary Region and up to five secondary Regions.
C.Deploy Amazon RDS for MySQL with cross-Region read replicas.
D.Deploy Amazon RDS for PostgreSQL with Multi-AZ and cross-Region snapshot copy.
AnswerB

Aurora Global Database provides low-latency reads and managed failover with minimal data loss.

Why this answer

Amazon Aurora Global Database is designed for low-latency cross-Region replication and provides a managed failover capability that promotes a secondary Region to primary with a Recovery Point Objective (RPO) of typically less than 1 second. This configuration meets the requirement of minimal data loss during a regional failure because replication is done at the storage layer, not through asynchronous binlog replication, ensuring near-zero lag.

Exam trap

The trap here is that candidates often confuse cross-Region read replicas (which are asynchronous and can lose data) with Aurora Global Database's storage-level replication, which provides near-zero RPO and automated promotion capabilities.

How to eliminate wrong answers

Option A is wrong because Amazon Aurora Serverless does not support cross-Region replication or Global Database; it is designed for intermittent workloads and scales automatically within a single Region. Option C is wrong because Amazon RDS for MySQL cross-Region read replicas use asynchronous replication with potentially higher lag and no automated promotion to primary, leading to greater data loss. Option D is wrong because Amazon RDS for PostgreSQL Multi-AZ provides high availability within a single Region, and cross-Region snapshot copy is not a replication mechanism; it requires manual restore and cannot achieve minimal data loss during failover.

460
MCQhard

An e-commerce application stores order data in Amazon RDS for MySQL. The database has grown to 1.5 TB and the company needs to retain data for 7 years for compliance. Current queries are becoming slow due to the large table size. The compliance requirement mandates that data older than 1 year must be retained but is rarely accessed. What strategy would reduce the active table size while maintaining compliance?

A.Create a read replica and run reports against it.
B.Partition the table by date and archive partitions older than 1 year to Amazon S3 using AWS DMS.
C.Delete data older than 1 year and use automated backups for compliance.
D.Vertically partition the table to separate frequently and infrequently accessed columns.
AnswerB

Removes old data from active table, retains in S3 for compliance.

Why this answer

Partitioning the table by date allows you to efficiently archive older, rarely accessed data to Amazon S3 using AWS DMS, reducing the active table size while retaining data for 7 years as required. This approach maintains compliance by keeping the archived data accessible in S3, and it improves query performance on the active partition by reducing the volume of data scanned.

Exam trap

The trap here is that candidates may think deleting old data and relying on backups is sufficient for compliance, not realizing that automated backups are for disaster recovery, not long-term retention of deleted records, and that partitioning with archival to S3 is the only option that both reduces active table size and meets the 7-year retention mandate.

How to eliminate wrong answers

Option A is wrong because creating a read replica does not reduce the active table size; it only offloads reporting queries, but the replica still contains the full 1.5 TB dataset, so slow queries due to large table size persist. Option C is wrong because deleting data older than 1 year violates the compliance requirement to retain data for 7 years; automated backups are for point-in-time recovery, not for long-term archival of deleted records. Option D is wrong because vertical partitioning (splitting columns) does not address the issue of large row counts; it only separates columns, leaving the number of rows unchanged, so query performance on the large table remains degraded.

461
MCQeasy

A company has a 100 GB MySQL database on an EC2 instance. They want to migrate to Amazon RDS for MySQL with minimal downtime. They have set up replication from the source to the target using MySQL native replication. After enabling replication, the 'Seconds_Behind_Master' value is increasing. The source database is write-heavy. What should the team do to reduce replication lag?

A.Tune the source database to reduce write load.
B.Enable Multi-AZ on the RDS instance.
C.Increase the RDS instance size to a larger instance class.
D.Switch to AWS DMS for migration.
AnswerC

Increasing the RDS instance size gives more CPU and memory resources to the database, enabling it to apply replication changes faster and reduce the 'Seconds_Behind_Master' value.

Why this answer

The replication lag is increasing because the RDS instance is not powerful enough to apply changes from the write-heavy source as fast as they arrive. Increasing the RDS instance class (option C) provides more CPU and memory resources, allowing the replica to apply transactions more quickly and reduce lag. Option A (tuning the source) might help but is not the most direct solution if the target is the bottleneck.

Option B (Multi-AZ) improves availability, not replication performance. Option D (AWS DMS) is an alternative migration tool but does not directly address lag in an existing native replication setup.

462
Multi-Selecthard

A company is using Amazon DynamoDB with provisioned capacity. The application is experiencing throttling on write requests. The database specialist needs to identify the cause. Which TWO metrics should be reviewed in CloudWatch? (Select TWO.)

Select 2 answers
A.ConsumedWriteCapacityUnits
B.WriteThrottleEvents
C.ThrottledWriteRequests
D.ReadThrottleEvents
E.SuccessfulRequestLatency
AnswersA, B

ConsumedWriteCapacityUnits shows the write capacity used; if it approaches provisioned capacity, throttling may occur.

Why this answer

'ConsumedWriteCapacityUnits' shows the actual write capacity used, helping to determine if provisioned capacity is exceeded, which leads to throttling. Option B is correct because 'WriteThrottleEvents' directly indicates the number of throttled write requests. Option C is incorrect because 'ThrottledWriteRequests' is not a valid CloudWatch metric for DynamoDB; the correct metric for throttled writes is 'WriteThrottleEvents'.

Option D is incorrect because 'ReadThrottleEvents' pertains to read throttling, not writes. Option E is incorrect because 'SuccessfulRequestLatency' measures latency, not throttling events.

463
MCQeasy

A company wants to encrypt an existing unencrypted Amazon RDS for PostgreSQL DB instance. What is the correct procedure?

A.Take a snapshot of the instance, create an encrypted copy of the snapshot, and restore the encrypted snapshot to a new DB instance.
B.Take a snapshot of the instance and restore it with encryption enabled.
C.Modify the DB instance and enable encryption in the RDS console.
D.Create a read replica of the instance and enable encryption on the replica.
AnswerA

This is the standard procedure to migrate to an encrypted instance.

Why this answer

Encryption for an existing unencrypted Amazon RDS for PostgreSQL DB instance cannot be enabled directly. The correct procedure is to take a snapshot of the instance, create an encrypted copy of that snapshot, and then restore the encrypted snapshot to a new DB instance. Option A accurately describes this process.

Option B is incorrect because restoring a snapshot does not allow enabling encryption during the restore; encryption must be applied at the time of snapshot copy. Option C is incorrect because you cannot modify a running DB instance to enable encryption. Option D is incorrect because creating a read replica does not encrypt the primary instance; encryption must be set up before replica creation.

464
MCQmedium

A company is running an Amazon Aurora MySQL database cluster. The database specialist notices that the write latency is high during peak hours. The cluster consists of one writer and two reader instances. Which action should the specialist take to reduce write latency?

A.Enable Auto Scaling on the cluster to automatically adjust capacity.
B.Increase the instance class of the writer instance to a larger size.
C.Enable Multi-AZ deployment for the cluster.
D.Add more reader instances to distribute the read load.
AnswerB

A larger instance class provides more CPU and memory, reducing write latency.

Why this answer

Increasing the instance class of the writer instance provides more CPU and memory resources, which can directly improve write throughput and reduce write latency during peak loads. Option A is incorrect because Auto Scaling in Aurora adjusts the number of reader instances, not the writer capacity. Option C is incorrect because Multi-AZ is already inherent in Aurora; enabling Multi-AZ does not affect write latency.

Option D is incorrect because adding more reader instances distributes read traffic but does not reduce write latency on the writer.

465
MCQhard

A company has an Amazon DynamoDB table with on-demand capacity mode. They notice that write requests are being throttled during peak hours. The table has a global secondary index (GSI) that is also throttled. Which action should the database specialist take to resolve the throttling?

A.Review the partition key design and consider adding a suffix to distribute writes more evenly.
B.Enable DynamoDB Streams to offload write operations.
C.Switch to provisioned capacity mode and increase write capacity units (WCU).
D.Increase the write capacity of the GSI by updating the table's provisioned throughput.
AnswerA

Even distribution of write traffic across partitions reduces throttling.

Why this answer

The throttling of both the base table and the GSI during peak hours indicates a hot partition caused by an uneven distribution of write activity. Reviewing the partition key design and adding a suffix to distribute writes more evenly is the correct action because it addresses the root cause: DynamoDB's on-demand mode automatically scales capacity, but it cannot overcome a skewed access pattern that overloads a single partition. By spreading writes across more partitions, you eliminate the hot spot and prevent throttling without changing capacity mode.

Exam trap

The trap here is that candidates assume throttling is always a capacity issue and choose to increase provisioned throughput, but the real problem is a hot partition caused by an uneven write pattern, which on-demand mode cannot automatically resolve.

How to eliminate wrong answers

Option B is wrong because DynamoDB Streams capture a time-ordered sequence of item-level changes but do not offload write operations; they are used for event-driven processing, not for reducing write load. Option C is wrong because switching to provisioned capacity mode and increasing WCU does not fix the underlying hot partition issue; throttling will persist if writes are concentrated on a few partitions, and on-demand mode already provides unlimited throughput per partition. Option D is wrong because you cannot directly increase the write capacity of a GSI in on-demand mode; GSI throughput is shared with the base table, and throttling occurs when a GSI partition is overloaded due to uneven write distribution, not because of insufficient provisioned capacity.

466
MCQeasy

A company is designing a document management system using Amazon S3 and needs to store metadata such as document ID, owner, creation date, and tags. The metadata must be searchable with low latency, supporting queries like 'Find all documents owned by user X with tag Y created after date Z'. Which AWS database service is most suitable for storing and querying this metadata?

A.Amazon DynamoDB with a GSI on (owner, creation_date) and a filter on tags.
B.Amazon Redshift Spectrum querying metadata stored in S3 as CSV.
C.Amazon RDS for PostgreSQL with a normalized schema.
D.Amazon ElastiCache for Redis with sorted sets for tags.
AnswerA

DynamoDB provides fast queries and flexible indexing.

Why this answer

Amazon DynamoDB is the most suitable choice because it provides single-digit millisecond latency for queries at any scale, which meets the low-latency search requirement. By creating a Global Secondary Index (GSI) on (owner, creation_date), you can efficiently query documents by owner and date range, and then apply a filter expression on tags to narrow results. This schema avoids the overhead of joins and normalization, making it ideal for high-throughput metadata lookups.

Exam trap

The trap here is that candidates often choose a relational database like PostgreSQL because they think normalized schemas are required for complex queries, but DynamoDB's GSI and filter expressions can handle this access pattern more efficiently at scale without the overhead of joins.

How to eliminate wrong answers

Option B is wrong because Amazon Redshift Spectrum is designed for analytical queries on large datasets in S3, not for low-latency, point-query or filtered lookups on metadata; it incurs significant overhead for each query and does not support sub-second response times. Option C is wrong because Amazon RDS for PostgreSQL with a normalized schema would require complex joins and indexing to support the multi-condition query, and relational databases typically have higher latency and scaling limitations compared to DynamoDB for this access pattern. Option D is wrong because Amazon ElastiCache for Redis with sorted sets is an in-memory cache, not a durable database; it lacks native support for multi-attribute queries like filtering by owner, date, and tags simultaneously, and sorted sets are optimized for leaderboard-style range queries, not arbitrary metadata searches.

467
Multi-Selecteasy

A company uses Amazon ElastiCache for Redis. They want to monitor cache hit ratio. Which TWO metrics should be used to calculate the cache hit ratio?

Select 2 answers
A.GetTypeCmds
B.CacheHits
C.Evictions
D.CacheMisses
E.CurItems
AnswersB, D

Number of successful key lookups.

Why this answer

Cache hit ratio is calculated as CacheHits / (CacheHits + CacheMisses). Therefore, the two metrics needed are CacheHits (option B) and CacheMisses (option D). Options A, C, and E are not used in this calculation: GetTypeCmds counts total get commands, Evictions indicates memory pressure, and CurItems shows the number of items in the cache.

468
MCQeasy

A company is migrating a MySQL database from on-premises to Amazon RDS for MySQL. The current database has several stored procedures and triggers that use user-defined functions (UDFs) compiled as shared libraries. What is the best practice for handling these UDFs in RDS?

A.Use Amazon RDS Custom for MySQL to upload the UDF libraries.
B.Use AWS Lambda to replace the UDFs.
C.Migrate to Amazon Aurora MySQL, which supports custom UDFs.
D.Refactor the stored procedures to avoid using the custom UDFs.
AnswerD

RDS does not support custom compiled UDFs; the application must be refactored.

Why this answer

Amazon RDS for MySQL does not allow access to the underlying file system, so you cannot upload custom UDF shared libraries (.so files). The best practice is to refactor the stored procedures and triggers to remove dependencies on these UDFs, replacing their logic with native MySQL functions or application-level code. This ensures compatibility with the managed RDS environment without requiring custom binaries.

Exam trap

The trap here is that candidates assume RDS Custom or Aurora MySQL will support custom UDFs, but neither service allows loading arbitrary shared libraries, making refactoring the only viable option.

How to eliminate wrong answers

Option A is wrong because Amazon RDS Custom for MySQL still restricts custom UDFs; RDS Custom provides OS-level access for patching and configuration but does not support loading arbitrary shared libraries for UDFs. Option B is wrong because AWS Lambda is an event-driven compute service that cannot directly replace UDFs used inside stored procedures or triggers; it would require significant architectural changes and introduce latency. Option C is wrong because Amazon Aurora MySQL does not support custom UDFs compiled as shared libraries; it only supports a limited set of built-in functions and Lambda-based functions via the native function interface.

469
MCQeasy

A company runs a reporting application on Amazon Redshift. The application queries a large fact table that is distributed by a key. The report queries filter on a date column. The report performance is slow. The database has 10 nodes. The company wants to improve query performance by optimizing the table design. Which design change should be made?

A.Set the sort key to the date column.
B.Increase the number of nodes in the cluster.
C.Change the distribution style to ALL to avoid data redistribution.
D.Change the distribution style to KEY on the date column.
AnswerA

Sort keys enable efficient range filtering, improving query performance for date-based filters.

Why this answer

Setting the sort key to the date column improves query performance by enabling range-restricted scans. When queries filter on a date column, Redshift uses zone maps to skip blocks that do not contain relevant data, drastically reducing the number of rows scanned. This is the most direct and cost-effective optimization for filter-heavy workloads on large fact tables.

Exam trap

The trap here is that candidates often confuse the purpose of distribution keys (for join co-location) with sort keys (for filter pruning), leading them to choose distribution changes (options C or D) instead of the correct sort key optimization.

How to eliminate wrong answers

Option B is wrong because increasing the number of nodes adds compute and storage capacity but does not address the root cause of slow scans; it is a scale-up solution that incurs additional cost without optimizing data access patterns. Option C is wrong because changing the distribution style to ALL replicates the entire table to every node, which eliminates data redistribution for joins but does not improve the efficiency of range-restricted scans on the date column; it also wastes storage and can degrade load performance. Option D is wrong because changing the distribution style to KEY on the date column would distribute rows based on date values, which can cause data skew if the date column has uneven cardinality (e.g., recent dates dominating), and it does not enable the block-minimax pruning that a sort key provides.

470
MCQmedium

A company is running an Amazon RDS for PostgreSQL DB instance with Multi-AZ deployment. They notice that the primary DB instance is experiencing high CPU utilization. The read replica shows normal CPU. Which action should the DBA take to reduce the load on the primary instance?

A.Failover to the standby instance
B.Increase the DB instance size
C.Convert the read replica to a Multi-AZ standby
D.Offload SELECT queries to the read replica
AnswerD

Offloading SELECT queries to the read replica directly reduces CPU utilization on the primary because read replicas can handle read-only traffic, leaving the primary to process write operations.

Why this answer

Offloading SELECT queries to the read replica reduces CPU utilization on the primary instance because read replicas can handle read traffic independently. Option A is incorrect: failing over to the standby instance does not reduce CPU load—the standby is only for high availability and becomes the new primary, still handling the same workload. Option B is incorrect: increasing the DB instance size would help but is not the best first step; it involves scaling costs and potential downtime, whereas using a read replica is more efficient and cost-effective for read-heavy workloads.

Option C is incorrect: converting the read replica to a Multi-AZ standby would make it a synchronous replica for failover, not for offloading reads, so it would not reduce CPU on the primary.

Exam trap

Candidates often confuse Multi-AZ standby with read replicas. The standby is only for high availability and does not serve read traffic, so failing over does not reduce CPU load.

471
MCQhard

A company is migrating an on-premises Oracle database to Amazon RDS for Oracle. The database size is 2 TB and the network bandwidth is limited. The company needs to minimize downtime during migration. Which AWS service should be used?

A.Oracle Data Pump export and import
B.AWS Schema Conversion Tool (SCT)
C.AWS Database Migration Service (DMS)
D.AWS Snowball Edge
AnswerC

DMS can migrate live data with minimal downtime.

Why this answer

AWS DMS is the correct choice because it supports ongoing replication from Oracle to Amazon RDS for Oracle, allowing you to perform a live migration with minimal downtime. With a 2 TB database and limited bandwidth, DMS can use Change Data Capture (CDC) to keep the target synchronized while the initial full load runs, then switch over with only a brief outage. This makes it ideal for minimizing downtime compared to offline methods.

Exam trap

The trap here is that candidates often choose AWS Snowball Edge for any large database with limited bandwidth, overlooking that DMS with CDC can achieve minimal downtime even for multi-terabyte databases, whereas Snowball introduces significant operational delay and is better suited for petabyte-scale or disconnected environments.

How to eliminate wrong answers

Option A is wrong because Oracle Data Pump export and import is an offline, batch-oriented tool that requires the source database to be quiesced or taken offline during the export, causing significant downtime, and it does not handle ongoing replication. Option B is wrong because AWS Schema Conversion Tool (SCT) is designed for converting database schemas and code between different database engines (e.g., Oracle to Aurora), not for migrating data with minimal downtime; it is often used alongside DMS but cannot perform the data transfer itself. Option D is wrong because AWS Snowball Edge is a physical data transfer device intended for large-scale offline migrations (typically 10 TB+) when network bandwidth is extremely limited, but it introduces weeks of shipping and manual handling, which increases downtime and complexity for a 2 TB database that could be migrated online with DMS.

472
Multi-Selecthard

A company is using Amazon DynamoDB for a gaming leaderboard that updates frequently. They need to maintain a sorted list of top 100 players by score. Which THREE design patterns can achieve this efficiently?

Select 3 answers
A.Use a Global Secondary Index (GSI) with score as the sort key and query with ScanIndexForward=false and Limit=100.
B.Use DynamoDB Accelerator (DAX) to cache query results.
C.Use DynamoDB Streams and AWS Lambda to maintain a separate leaderboard table with the top 100 scores.
D.Scan the entire table and sort the results in memory.
E.Use Amazon ElastiCache for Redis with sorted sets to maintain the leaderboard.
AnswersA, C, E

This retrieves the top 100 scores efficiently.

Why this answer

A Global Secondary Index (GSI) with score as the sort key allows you to query items in descending order using ScanIndexForward=false and limit the result to the top 100 players. This pattern efficiently retrieves the highest scores without scanning the entire table, leveraging DynamoDB's index query capabilities.

Exam trap

The DBS-C01 exam often tests the misconception that DAX can perform sorting or ranking operations, but DAX is only a cache and cannot reorder data or maintain sorted sets.

473
MCQhard

A company is running a production PostgreSQL database on an EC2 instance (db.m5.xlarge) with 500 GB of gp2 EBS storage. The database is used by a customer-facing application that requires low latency. The company plans to migrate this database to Amazon RDS for PostgreSQL with minimal downtime. The current database has a high write load with frequent updates and deletes, and the table sizes are growing. The company also wants to enable Multi-AZ for high availability and use read replicas for reporting workloads. During migration planning, they discover that the source database has several large unlogged tables and uses custom PostgreSQL extensions that are not available in RDS. Which migration strategy should the company use to minimize downtime and meet all requirements?

A.Use AWS DMS with ongoing replication, convert unlogged tables to logged tables, and migrate custom extensions using AWS SCT.
B.Use pg_dump to export the database and pg_restore to import into RDS, then set up read replicas.
C.Copy the database files to Amazon S3, then use the rdsadmin.rdsadmin_restore_from_s3 procedure to restore to RDS.
D.Set up PostgreSQL streaming replication from the EC2 instance to an RDS read replica, then promote the replica.
AnswerA

DMS can handle unlogged tables by replicating data as regular tables, and SCT can suggest alternatives for unsupported extensions.

Why this answer

AWS DMS with ongoing replication allows near-zero downtime by continuously replicating changes from the source to the target RDS instance. DMS can handle unlogged tables by converting them to logged tables during migration, as DMS requires logical replication which relies on write-ahead logs. AWS SCT can help assess and convert custom PostgreSQL extensions to RDS-compatible equivalents or suggest alternatives.

Option B (pg_dump/pg_restore) is an offline approach that would cause significant downtime, which does not meet the minimal downtime requirement. Option C (copy to S3 then restore) is not a supported migration method for PostgreSQL to RDS. Option D (streaming replication) is not feasible because RDS does not accept direct streaming replication from an external source, and unlogged tables cannot be replicated via streaming.

Therefore, A is the best strategy.

474
MCQhard

A database team uses Amazon DynamoDB with auto scaling enabled. They observe frequent throttling on a table during peak hours. The table's read capacity is set to 5000 RCU with auto scaling range 3000-7000. The consumed read capacity graph shows spikes to 6000 RCU but throttling occurs at 5500. What is the most likely cause?

A.Write capacity units are insufficient
B.Auto scaling is disabled for the table
C.The table has too many partitions
D.Auto scaling cannot react quickly enough to sudden traffic spikes
AnswerD

Auto scaling has a lag; spikes can exceed provisioned capacity before scaling completes.

Why this answer

Auto scaling uses a target utilization (default 70%) and cannot scale fast enough for sudden spikes. Option A is wrong because auto scaling is enabled. Option B is wrong because WCU are separate.

Option C is wrong because partition count doesn't directly cause throttling if RCU is sufficient.

475
MCQhard

A company runs a production Amazon RDS for PostgreSQL Multi-AZ DB instance (db.r5.large) with 500 GB of General Purpose SSD (gp2) storage. The application experiences intermittent latency spikes every 15 minutes. Monitoring shows that during these spikes, the ReadIOPS metric on the primary instance spikes to 5,000 IOPS (the baseline is 1,500 IOPS), and the BurstBalance drops from 100% to 20% then recovers. There is no increase in CPU or connections. The application uses connection pooling with pgBouncer on an EC2 instance. The team has verified that no long-running queries or index scans are causing the spikes. Which action is MOST likely to resolve the intermittent latency?

A.Create a read replica and redirect read traffic to it.
B.Increase the DB instance to db.r5.xlarge to improve CPU and network performance.
C.Migrate the storage to gp3 with a baseline of 3,000 IOPS and 125 MB/s throughput.
D.Scale the storage to 1,000 GB to increase baseline IOPS and burst credits.
AnswerC

gp3 provides consistent baseline IOPS without burst credits, eliminating the performance variability due to credit exhaustion.

Why this answer

The latency spikes are caused by gp2 storage burst credit exhaustion. The 500 GB gp2 volume has a baseline of 1,500 IOPS, but the workload spikes to 5,000 IOPS every 15 minutes, rapidly consuming burst credits. Migrating to gp3 provides a baseline of 3,000 IOPS and 125 MB/s throughput without relying on burst credits, thus eliminating the credit exhaustion issue.

Option A (read replica) does not resolve the primary instance's write IOPS spikes. Option B (larger instance) does not address storage IOPS limitations; CPU and connections are already normal. Option D (scale storage to 1,000 GB) would increase the gp2 baseline to 3,000 IOPS and provide more burst credits, but gp3 offers a simpler, more cost-effective solution with consistent performance and no credit-based throttling.

Exam trap

Candidates often confuse gp2 burst credits with gp3's fixed performance. They may think increasing volume size alone will eliminate bursts, but gp2 always uses credits for spikes above baseline. Migrating to gp3 removes the burst mechanism entirely.

476
MCQeasy

A company wants to deploy an Amazon RDS for MySQL database for a new application. The database must be highly available with automatic failover. Which configuration should they choose?

A.Deploy a single-AZ instance with automated backups
B.Deploy a cross-region replica
C.Deploy a single-AZ instance with a read replica
D.Deploy a Multi-AZ instance with a standby replica
AnswerD

Multi-AZ provides automatic failover to the standby.

Why this answer

A Multi-AZ deployment for Amazon RDS automatically provisions and maintains a synchronous standby replica in a different Availability Zone. In the event of a failure, Amazon RDS automatically fails over to the standby, providing high availability without manual intervention. This configuration meets the requirement for automatic failover and high availability.

Exam trap

The trap here is that candidates often confuse read replicas with Multi-AZ standby replicas, assuming that a read replica can provide automatic failover, but read replicas require manual promotion and may have data loss due to asynchronous replication.

How to eliminate wrong answers

Option A is wrong because a single-AZ instance with automated backups provides point-in-time recovery but does not offer automatic failover; if the instance fails, downtime occurs until a new instance is restored from backups. Option B is wrong because a cross-region replica is an asynchronous read replica used for disaster recovery or read scaling, not for automatic failover within the same region; it requires manual promotion to become the primary. Option C is wrong because a single-AZ instance with a read replica provides read scaling and can be manually promoted for failover, but it does not provide automatic failover; the read replica is asynchronous and may have replication lag.

477
MCQhard

A multinational e-commerce company runs an Amazon Aurora MySQL database for its product catalog. The database is 2 TB and has a high write volume. The company needs to create a test environment that contains a subset (10%) of the production data for developers to use. The test environment must be refreshed daily with the latest production data. The operations team wants to minimize cost and ensure that the test environment does not impact production performance. Which solution should they implement?

A.Use AWS DMS to continuously replicate a filtered subset of data to a test cluster.
B.Take a manual snapshot of the production cluster, restore it to a new cluster, and delete 90% of the data.
C.Use Aurora cloning to create a clone of the production cluster. Use database triggers or scripts to delete 90% of the data after cloning.
D.Create a read replica of the production cluster, promote it to a standalone cluster, and delete 90% of the data.
AnswerC

Cloning is fast, cost-effective, and does not impact production.

Why this answer

Using Aurora cloning creates a storage-level copy that shares the same underlying storage as the source cluster, minimizing cost and avoiding any performance impact on production because cloning does not involve copying data. After cloning, you can delete 90% of the data using scripts or triggers to retain only 10% for testing. Option A (DMS continuous replication) would incur additional cost and ongoing replication overhead.

Option B (snapshot and restore) would create a full 2 TB copy and then you would have to delete data, incurring storage costs for the full copy. Option D (read replica) would impact production replication and would also require deleting data after promotion.

478
MCQhard

A database administrator has the IAM policy shown in the exhibit. Which action will be allowed by this policy?

A.Modify the prod-db instance.
B.Create a snapshot of the prod-db instance.
C.Delete the prod-db instance.
D.Describe all DB instances in the account.
AnswerB

Explicitly allowed.

Why this answer

The policy explicitly allows CreateDBSnapshot on the resource. Option A is denied. Option C is not in the policy.

Option D is not in the policy.

479
MCQhard

A company runs a large-scale e-commerce platform using Amazon RDS for MySQL with a Multi-AZ deployment. The database has a table 'orders' with 200 million rows. Recently, they added a new index on the 'order_date' column to improve reporting queries. After adding the index, they noticed increased write latency and occasional replication lag. The application writes new orders continuously. The table experiences about 10,000 writes per second. The DB instance is db.r5.4xlarge. The index creation was done using the ALTER TABLE statement with a default algorithm. What is the most likely cause of the increased write latency and replication lag?

A.The index creation DDL statement is not replicated to the standby instance, causing inconsistency.
B.The instance size is insufficient for the write workload.
C.The index was created using the default algorithm (COPY), which locks the table and blocks writes, causing replication lag.
D.The new index is causing excessive overhead on write operations due to index maintenance.
AnswerC

In MySQL 5.6 and 5.7, ALTER TABLE uses COPY algorithm by default, which locks the table for writes during the operation.

Why this answer

The default algorithm for ALTER TABLE in MySQL is COPY, which creates a new table, copies all rows, and rebuilds indexes. During this process, the table is locked with a write lock, blocking DML operations and causing increased write latency. In a Multi-AZ deployment, the DDL is replicated to the standby, but the lock on the primary delays writes, which can manifest as replication lag when the standby applies the same blocking DDL.

Exam trap

The trap here is that candidates often assume any index addition causes permanent write overhead (Option D), but the question describes a sudden latency spike immediately after the operation, which is characteristic of the blocking COPY algorithm, not ongoing maintenance.

How to eliminate wrong answers

Option A is wrong because DDL statements like ALTER TABLE are replicated to the standby instance via the binary log in MySQL Multi-AZ deployments; the index creation is not skipped, so inconsistency does not occur. Option B is wrong because the db.r5.4xlarge instance (16 vCPUs, 128 GB memory) is more than sufficient for 10,000 writes per second on a single table; the issue is not raw capacity but the blocking nature of the DDL operation. Option D is wrong because while index maintenance does add overhead to writes, the sudden increase in write latency and replication lag immediately after adding the index points to the blocking DDL operation itself, not the ongoing maintenance cost of the new index.

480
MCQeasy

A company has an Amazon RDS for PostgreSQL DB instance with automated backups enabled. The retention period is set to 7 days. A developer accidentally performed a DROP TABLE operation on a critical table 2 days ago. How can the table be recovered with minimal data loss?

A.Perform a point-in-time restore to a time just before the DROP TABLE operation.
B.Use the pg_dump utility to create a manual backup and restore it.
C.Create a read replica of the DB instance and promote it to a standalone instance.
D.Restore the DB instance from the oldest automated snapshot.
AnswerA

PITR allows recovery to any point within the retention period.

Why this answer

The correct approach is to perform a point-in-time restore (PITR) to a time just before the DROP TABLE operation. RDS for PostgreSQL with automated backups enabled supports PITR within the backup retention period (7 days). Since the table was dropped 2 days ago, restoring the DB instance to a point in time just before the drop will recover the table with minimal data loss.

Option B (pg_dump) is not a suitable recovery method because it requires a pre-existing backup; it cannot generate a backup retroactively. Option C (creating a read replica and promoting it) would create a copy of the current state, which already lacks the dropped table, so it does not help. Option D (restore from oldest automated snapshot) would revert the entire database to a 7-day-old state, losing all changes made in the last 7 days, including the table that was dropped only 2 days ago, resulting in greater data loss.

481
MCQeasy

A company is using Amazon DynamoDB for a web application. The company notices that read requests to a particular table are throttled during peak hours. The table has a provisioned read capacity of 1000 read capacity units (RCUs). The read requests are mostly eventually consistent reads. What should the company do to reduce throttling without changing the application code?

A.Use Amazon ElastiCache to cache the read results.
B.Switch to strongly consistent reads to reduce the number of read requests.
C.Implement DynamoDB Accelerator (DAX) to cache read requests.
D.Increase the provisioned read capacity for the table.
AnswerD

Increasing provisioned read capacity directly addresses the throttling by allowing more read requests per second without any code changes.

Why this answer

Increasing provisioned read capacity directly provides more RCUs, reducing throttling without code changes. Option A is wrong because ElastiCache would require application code changes to implement caching. Option B is wrong: Switching to strongly consistent reads would double RCU consumption per read (for reads up to 4 KB) or increase it, worsening throttling.

Option C is wrong: DAX does require code changes to integrate (though minimal), and the question specifically asks for no application code changes.

482
MCQhard

A company is using Amazon DynamoDB for a high-traffic application. The application is experiencing intermittent `ProvisionedThroughputExceededException` errors. The team has already increased the read and write capacity units multiple times but the errors persist. Which of the following is the MOST likely cause of the issue?

A.DynamoDB Accelerator (DAX) is not properly configured
B.The table is part of a DynamoDB Global Table and replication is causing conflicts
C.The provisioned capacity is not increased enough
D.A hot key or uneven partition access pattern is causing throttling
AnswerD

Hot keys or uneven partition access patterns can cause throttling even if overall provisioned capacity appears sufficient.

Why this answer

Hot keys or uneven partition access patterns can cause throttling even if overall provisioned capacity appears sufficient. Option A is wrong because DynamoDB Accelerator (DAX) is an in-memory cache that reduces read load, but it does not directly fix capacity exceeded errors; misconfiguration might cause cache misses but not ProvisionedThroughputExceededException. Option B is wrong because Global Tables replication does not cause throttling on the source table; conflicts are handled by last-writer-wins.

Option C is wrong because increasing capacity units multiple times without resolving the underlying access pattern suggests the issue is not simply insufficient capacity.

483
MCQmedium

A company is migrating a 200 GB PostgreSQL database to Amazon Aurora PostgreSQL. They want to use the AWS DMS console to create a migration task. The source database is in a VPC and the target Aurora cluster is in the same VPC. The DMS replication instance is also in the same VPC. The source database is publicly accessible. What additional configuration is required to enable connectivity between DMS and the source database?

A.Associate the DMS replication instance with a subnet group that includes public subnets.
B.Place the source database in the same public subnet as the DMS replication instance.
C.Create a VPC endpoint for the source database service.
D.Configure the source database's security group to allow inbound traffic from the DMS replication instance's security group.
AnswerD

The security group must allow traffic from DMS to reach the database.

Why this answer

The source database is publicly accessible but resides in a VPC, and the DMS replication instance is in the same VPC. To allow the DMS replication instance to connect to the source database, the source database's security group must have an inbound rule that permits traffic on the database port (e.g., 5432 for PostgreSQL) from the DMS replication instance's security group. This ensures that only the replication instance can initiate the connection, maintaining security within the VPC.

Exam trap

The trap here is that candidates assume a publicly accessible source database requires public subnet placement or a VPC endpoint, but the key is that both resources are in the same VPC, so security group rules alone enable connectivity without exposing the database to the internet.

How to eliminate wrong answers

Option A is wrong because associating the DMS replication instance with a subnet group that includes public subnets is unnecessary and irrelevant; the replication instance is already in the same VPC as the source and target, and connectivity is controlled by security groups, not subnet type. Option B is wrong because placing the source database in the same public subnet as the DMS replication instance is not required and may violate security best practices; the source database can remain in its current subnet as long as security group rules allow traffic. Option C is wrong because a VPC endpoint is used to privately connect to AWS services (e.g., S3, DynamoDB) without traversing the internet, but the source database is a self-managed PostgreSQL instance, not an AWS service, so a VPC endpoint does not apply.

484
MCQmedium

A company runs a customer relationship management (CRM) application on Amazon RDS for PostgreSQL. The application stores customer data in a table with over 50 million rows. The company recently added a new query that searches for customers by their email domain (e.g., '@example.com'). The query uses a LIKE pattern: 'WHERE email LIKE ''%@example.com'''. The query takes over 30 seconds to complete. The DBA has already created a B-tree index on the email column, but it does not help. Which action should the database specialist recommend to improve query performance?

A.Create a hash index on the email column.
B.Increase the shared_buffers parameter to improve caching.
C.Create a B-tree index on the reversed email string.
D.Create a trigram index (using pg_trgm extension) on the email column.
AnswerD

Trigram indexes are designed for fast LIKE queries.

Why this answer

The query uses a leading wildcard LIKE pattern ('%@example.com'), which prevents a standard B-tree index from being used because the search string does not have a fixed prefix. A trigram index, provided by the pg_trgm extension, breaks strings into three-character substrings (trigrams) and allows the database to efficiently match patterns with leading wildcards. This index type is specifically designed for fuzzy text matching and LIKE queries with wildcards, reducing the query time from over 30 seconds to milliseconds.

Exam trap

The trap here is that candidates assume a B-tree index can handle all LIKE patterns, but AWS specifically tests the understanding that leading wildcards disable B-tree index scans, requiring a specialized index like pg_trgm for pattern-matching performance.

How to eliminate wrong answers

Option A is wrong because hash indexes in PostgreSQL only support equality comparisons (=), not pattern-matching operations like LIKE. Option B is wrong because increasing shared_buffers improves caching of data pages but does not change the query execution plan; the B-tree index is still not used for leading-wildcard searches, so the query remains a full table scan. Option C is wrong because creating a B-tree index on the reversed email string would only help if the query were rewritten to use a trailing wildcard (e.g., WHERE REVERSE(email) LIKE 'moc.elpmaxe@%'), which is not the given query pattern and adds complexity without addressing the leading wildcard issue.

485
MCQeasy

A developer notices that an Amazon ElastiCache for Redis cluster is experiencing high latency. The cluster uses a single node. Which CloudWatch metric should be reviewed first to determine if the issue is due to memory pressure?

A.NetworkBytesIn
B.ReplicationLag
C.CPUUtilization
D.DatabaseMemoryUsagePercentage
AnswerD

DatabaseMemoryUsagePercentage shows the percentage of the node's memory used. High usage can lead to memory pressure, causing latency due to evictions or swap.

Why this answer

The correct metric to check for memory pressure is DatabaseMemoryUsagePercentage. This metric shows the percentage of the node's available memory that is in use, and when high it can lead to latency due to evictions or swap usage. NetworkBytesIn measures network traffic, not memory.

ReplicationLag is relevant only for clusters with replicas, and CPUUtilization indicates CPU load, not memory pressure.

486
MCQmedium

A company is deploying a new multi-AZ Amazon RDS for PostgreSQL database. The security team requires that all traffic to the database be encrypted in transit. Which configuration ensures this?

A.Use a customer master key (CMK) for RDS
B.Enable SSL/TLS and require client connections to use SSL
C.Place the RDS instance in a public subnet
D.Enable encryption at rest using AWS KMS
AnswerB

SSL/TLS encrypts data in transit.

Why this answer

Enabling SSL/TLS on the RDS instance and requiring client connections to use SSL ensures that all data transmitted between clients and the database is encrypted in transit. This is achieved by setting the rds.force_ssl parameter to 1 in the DB parameter group, which enforces SSL/TLS for all connections, meeting the security team's requirement for encryption in transit.

Exam trap

The trap here is confusing encryption at rest (KMS, CMK) with encryption in transit (SSL/TLS), leading candidates to select options that secure data on disk but not during network transmission.

How to eliminate wrong answers

Option A is wrong because using a customer master key (CMK) for RDS encrypts data at rest, not in transit; CMKs are used with AWS KMS for storage encryption, not for network traffic. Option C is wrong because placing the RDS instance in a public subnet exposes it to the internet, which does not inherently encrypt traffic and violates security best practices; encryption in transit requires SSL/TLS, not subnet placement. Option D is wrong because enabling encryption at rest using AWS KMS protects data stored on disk, not data transmitted over the network; it does not address encryption in transit.

487
MCQeasy

A startup uses Amazon ElastiCache for Redis as a caching layer for its database. Users report that application responses are slow. The developer checks the ElastiCache metrics and sees that 'CacheHits' are low and 'CacheMisses' are high. What is the most likely cause?

A.The cluster does not have enough read replicas.
B.The ElastiCache cluster does not have enough write capacity.
C.The ElastiCache nodes have high CPU utilization.
D.The cache key TTL is too short, causing frequent evictions.
AnswerD

Short TTL leads to early eviction and cache misses.

Why this answer

A low cache hit ratio and high cache miss ratio indicate that the cache is not storing data that is frequently requested. The most likely cause is that the Time-To-Live (TTL) for cache keys is set too short, causing data to be evicted before it can be reused. Option A is incorrect because read replicas improve read scalability but do not directly affect cache hit ratio.

Option B is incorrect because write capacity is not relevant for a caching layer that primarily serves reads. Option C is incorrect: while high CPU utilization can cause latency, it would not specifically cause low cache hits and high misses.

488
Multi-Selecthard

A company uses Amazon Aurora MySQL-Compatible Edition. The security team wants to implement database activity streams to monitor database activity. Which THREE statements are true about Aurora database activity streams?

Select 3 answers
A.Activity streams can be started and stopped without restarting the database.
B.Activity streams are encrypted using a KMS key.
C.Activity streams automatically mask sensitive data in the logs.
D.Activity streams only capture DDL statements, not DML or SELECT.
E.Activity streams publish database activity to CloudWatch Logs and Kinesis Firehose.
AnswersA, B, E

Activity streams are started via the RDS console or API and do not require a restart.

Why this answer

Options A, B, and E are correct. Activity streams can be started and stopped without a database restart (A). They are encrypted using a KMS key (B).

They publish database activity to Amazon CloudWatch Logs and Amazon Kinesis Firehose (E). Option C is incorrect because activity streams do not automatically mask sensitive data; they capture the actual queries. Option D is incorrect because activity streams capture all SQL statements, including DDL, DML, and SELECT.

489
MCQmedium

Refer to the exhibit. A company is creating an Aurora MySQL cluster using the AWS CLI. The command fails with an error. The company has a default KMS key but the command specifies a customer-managed KMS key. What is the most likely cause of the failure?

A.The IAM user does not have permission to use the specified KMS key
B.The --kms-key-id parameter is not supported for aurora-mysql engine
C.The KMS key does not exist
D.The KMS key is in a different region
AnswerA

Permission to use the KMS key is required.

Why this answer

The most likely cause of the failure is that the IAM user does not have the required permissions to use the specified customer-managed KMS key. When you specify a KMS key in the `--kms-key-id` parameter for an Aurora MySQL cluster, the IAM user must have `kms:CreateGrant`, `kms:Decrypt`, and `kms:Encrypt` permissions on that key. If the user lacks these permissions, the command fails even if the key exists and is in the correct region.

Exam trap

The trap here is that candidates often assume the error is due to the key not existing or being in a different region, but the most common failure is an IAM permissions issue on the KMS key, especially when a customer-managed key is specified instead of the default AWS managed key.

How to eliminate wrong answers

Option B is wrong because the `--kms-key-id` parameter is fully supported for the `aurora-mysql` engine; Aurora MySQL supports encryption at rest using KMS keys. Option C is wrong because the question states the company has a default KMS key, and the command specifies a customer-managed KMS key, but the error is not about the key's existence—it's about permissions; if the key did not exist, the error would be 'Key not found' or similar. Option D is wrong because if the KMS key were in a different region, the CLI would return a 'Region mismatch' or 'InvalidKeyId' error, but the question does not indicate a cross-region scenario, and the most common cause in single-region setups is missing IAM permissions.

490
MCQmedium

A company runs an Amazon Redshift cluster with 8 dc2.large nodes for its data warehouse. The data engineering team loads data daily using COPY commands from S3. Recently, the load times have increased significantly. The cluster's CloudWatch metric 'CPUUtilization' is high during the load. The administrator runs the STL_LOAD_ERRORS table and finds no errors. The SVL_S3LOG shows that the COPY command is scanning many small files. The data in S3 is stored as 10,000 small CSV files (each ~100 KB). Which action will MOST improve the COPY performance?

A.Use the MANIFEST option to specify the files explicitly
B.Use the JSON format instead of CSV to reduce parsing overhead
C.Consolidate the small files into fewer, larger files (e.g., 100 files of 10 MB each)
D.Change the table's distribution style to ALL to avoid data redistribution
AnswerC

Larger files reduce the overhead of file opening and improve parallelism.

Why this answer

Consolidating many small files into fewer, larger files reduces the overhead of opening and processing numerous small files during the COPY command. Redshift performs better with larger files (e.g., 64 MB to 1 GB) because it can parallelize the load across slices more efficiently. Option A is incorrect because the MANIFEST option helps with specifying files but does not address the root cause of many small files.

Option B is incorrect because JSON format typically increases parsing overhead compared to CSV. Option D is incorrect because changing the distribution style to ALL does not improve COPY performance; it affects query performance after data is loaded.

Exam trap

Candidates may confuse the benefit of file format (JSON vs. CSV) with the performance impact of file size. The real issue is the large number of small files, not the format.

491
MCQhard

A financial services company is migrating a 10 TB Oracle data warehouse to Amazon Redshift. The source database uses a combination of partitioned tables, indexes, and materialized views. The migration team plans to use AWS SCT and DMS. Which approach should the team take to optimize query performance after migration?

A.Create indexes on frequently queried columns
B.Recreate all materialized views using standard views
C.Define appropriate distribution keys and sort keys
D.Partition tables by date using the PARTITION BY clause
AnswerC

These are the primary performance optimization mechanisms in Redshift.

Why this answer

Amazon Redshift is a columnar data warehouse that does not support traditional indexes or the PARTITION BY clause used in Oracle. Instead, query performance in Redshift is optimized by defining appropriate distribution keys (to minimize data shuffling across nodes) and sort keys (to enable zone maps and reduce the amount of data scanned). This approach directly addresses the performance needs of the migrated 10 TB data warehouse.

Exam trap

The trap here is that candidates familiar with traditional RDBMS concepts (indexes, table partitioning) assume they apply directly to Redshift, but Redshift uses a fundamentally different architecture where distribution and sort keys are the primary performance levers.

How to eliminate wrong answers

Option A is wrong because Amazon Redshift does not use traditional B-tree or bitmap indexes; it relies on sort keys and zone maps for data skipping. Option B is wrong because materialized views in Redshift can improve performance by pre-computing and storing results, whereas standard views execute the underlying query each time, often leading to slower performance. Option D is wrong because Redshift does not support the Oracle-style PARTITION BY clause; it uses distribution keys and sort keys to manage data layout across nodes.

492
Multi-Selecteasy

Which TWO AWS services can be used to migrate an on-premises SQL Server database to Amazon RDS for SQL Server?

Select 2 answers
A.AWS DataSync
B.AWS DMS
C.Native SQL Server backup to S3 and restore to RDS
D.AWS Storage Gateway
E.AWS DynamoDB
AnswersB, C

DMS supports SQL Server as source and target.

Why this answer

AWS DMS (Database Migration Service) is correct because it is purpose-built for migrating databases to AWS with minimal downtime, supporting homogeneous migrations like SQL Server to Amazon RDS for SQL Server. It uses change data capture (CDC) to replicate ongoing changes, enabling near-zero downtime migrations. Native SQL Server backup to S3 and restore to RDS is also correct because SQL Server's native backup/restore mechanism can be used to take a full backup, upload it to Amazon S3, and then restore it directly onto an RDS for SQL Server instance, which is a supported and common offline migration method.

Exam trap

The trap here is that candidates often assume only AWS DMS is valid for database migrations, overlooking the fact that native SQL Server backup/restore to S3 is a fully supported and often simpler offline migration method for SQL Server to RDS.

493
MCQeasy

A company is migrating an on-premises MongoDB database to Amazon DocumentDB. The migration must be online with minimal downtime. Which AWS service should be used?

A.AWS DataSync
B.AWS DMS
C.AWS S3
D.AWS Snowball Edge
AnswerB

DMS supports MongoDB to DocumentDB with CDC for minimal downtime.

Why this answer

AWS Database Migration Service (DMS) is the correct choice because it supports ongoing replication (change data capture) from MongoDB to Amazon DocumentDB, enabling an online migration with minimal downtime. DMS can perform a full load of existing data and then continuously replicate changes from the source MongoDB oplog to keep the target DocumentDB synchronized until cutover.

Exam trap

The trap here is that candidates often confuse AWS DataSync (file transfer) or Snowball Edge (offline bulk transfer) with database migration, but DMS is the only service that supports live, ongoing replication for heterogeneous database migrations like MongoDB to DocumentDB.

How to eliminate wrong answers

Option A is wrong because AWS DataSync is designed for moving large volumes of file data (e.g., NFS, SMB) to Amazon S3 or EFS, not for migrating databases with ongoing replication. Option C is wrong because Amazon S3 is an object storage service and cannot perform live database replication or schema conversion required for a MongoDB-to-DocumentDB migration. Option D is wrong because AWS Snowball Edge is a physical data transfer device for offline bulk data migration, which cannot support an online, minimal-downtime migration with continuous replication.

494
MCQmedium

A company uses Amazon Aurora MySQL for its customer relationship management (CRM) system. The database has a table "contacts" with millions of rows. The application frequently searches for contacts by email address. The email column has a B-tree index. The DBA notices that queries are still slow, and the EXPLAIN plan shows index scans but not index-only scans. What is the most likely cause?

A.The query selects columns not included in the index, requiring table lookups.
B.The index is a composite index on (email, phone) and the query selects only email.
C.The index has low cardinality.
D.The index type is not suitable for equality searches.
AnswerA

If the query selects columns like phone not in the index, the database must access the table, preventing an index-only scan.

Why this answer

An index-only scan requires that all columns referenced in the query (both in the SELECT list and WHERE clause) be present in the index. Since the query selects columns not included in the B-tree index on the email column, Aurora MySQL must perform additional table lookups (row fetches) to retrieve those missing columns, resulting in an index scan rather than an index-only scan.

Exam trap

The trap here is that candidates often assume any index scan is optimal, failing to recognize that an index-only scan (covered index) is significantly faster because it avoids table row lookups, and the EXPLAIN plan's 'Using index' vs. 'Using index condition' distinction is the key clue.

How to eliminate wrong answers

Option B is wrong because a composite index on (email, phone) would actually support index-only scans if the query selects only email, as the email column is the leading column and the index covers the query; the issue described is the opposite—missing columns in the index. Option C is wrong because low index cardinality (many duplicate values) would reduce the efficiency of index scans but would not prevent index-only scans; index-only scans are still possible as long as all required columns are in the index. Option D is wrong because a B-tree index is highly suitable for equality searches (e.g., WHERE email = '...'), and the EXPLAIN plan shows index scans, confirming the index is being used; the problem is not the index type but the need to fetch non-indexed columns.

495
MCQhard

A company needs to comply with PCI DSS requirements for an Amazon RDS for Oracle DB instance. The requirements include encryption of sensitive data at rest and in transit, and automated key rotation. Which combination of services and configurations should be used? (Select THREE.)

A.Use AWS CloudHSM to generate and store encryption keys.
B.Enable encryption at rest on the RDS instance using AWS KMS.
C.Use AWS Secrets Manager to automatically rotate database credentials.
D.Enable SSL/TLS for connections to the database.
E.Enable VPC Flow Logs to audit database connections.
AnswerB, C, D

Encryption at rest is required for PCI DSS.

Why this answer

Options B, C, and D are correct. Option B: Enable encryption at rest on the RDS instance using AWS KMS. Option C: Use AWS Secrets Manager to automatically rotate database credentials.

Option D: Enable SSL/TLS for connections to the database. Option A is incorrect because AWS CloudHSM is not required for key rotation; KMS can handle key rotation automatically. Option E is incorrect because VPC Flow Logs are for network traffic monitoring, not encryption.

496
MCQhard

A security team needs to grant an IAM user permission to modify only the 'db_secrets' secret in AWS Secrets Manager. Which IAM policy statement is correct?

A.{ 'Effect': 'Allow', 'Action': 'secretsmanager:UpdateSecret', 'Resource': '*' }
B.{ 'Effect': 'Allow', 'Action': 'secretsmanager:*', 'Resource': '*' }
C.{ 'Effect': 'Allow', 'Action': 'secretsmanager:PutSecretValue', 'Resource': 'arn:aws:secretsmanager:us-east-1:123456789012:secret:*' }
D.{ 'Effect': 'Allow', 'Action': 'secretsmanager:PutSecretValue', 'Resource': 'arn:aws:secretsmanager:us-east-1:123456789012:secret:db_secrets-??????' }
AnswerD

This restricts to the specific secret and allows PutSecretValue.

Why this answer

It grants the specific `secretsmanager:PutSecretValue` action on the exact ARN of the `db_secrets` secret, including the required six-character random suffix (`-??????`) that AWS Secrets Manager appends to secret names. This ensures the IAM user can only modify that single secret, meeting the security requirement of least privilege.

Exam trap

The trap here is that candidates often forget the mandatory six-character random suffix in Secrets Manager ARNs and use only the secret name, leading them to choose a wildcard resource like option C, which grants unintended access to multiple secrets.

How to eliminate wrong answers

Option A is wrong because it uses a wildcard resource (`'*'`), which would allow modifying any secret in the account, violating the requirement to restrict access to only `db_secrets`. Option B is wrong because it allows all Secrets Manager actions (`secretsmanager:*`) on all resources, granting far too broad permissions, including deleting or creating secrets. Option C is wrong because the resource ARN uses a wildcard (`'*'`) instead of the specific secret name with its random suffix, which would match multiple secrets and not restrict to `db_secrets` alone.

497
MCQhard

A company uses Amazon RDS for SQL Server with Multi-AZ deployment. The security team has mandated that all connections to the database must use SSL/TLS. The database is accessed by multiple applications running on EC2 instances. Which configuration ensures that all connections use SSL/TLS?

A.Modify the DB instance by enabling the 'Require SSL' option in the RDS console.
B.Set the parameter rds.force_ssl to 1 in the DB parameter group and revoke permissions from users that do not use SSL.
C.Set the parameter rds.force_ssl to 1 in the DB parameter group.
D.Configure the applications to use a certificate from a trusted certificate authority and connect using SSL.
AnswerC

Correct. Setting rds.force_ssl=1 in the DB parameter group enforces SSL/TLS for all connections to the RDS for SQL Server instance without any additional steps.

Why this answer

Setting the rds.force_ssl parameter to 1 in the DB parameter group enforces SSL/TLS for all connections to the RDS for SQL Server instance. No additional steps are needed; this parameter alone ensures that any connection attempt without SSL is rejected. Option B is incorrect because it includes an unnecessary step of revoking permissions from users that do not use SSL; the rds.force_ssl setting already forces SSL for all users and revoking permissions is redundant.

Option A is incorrect because the 'Require SSL' option does not exist in the RDS console; SSL enforcement is controlled via parameter group settings. Option D is incorrect because configuring applications to use a trusted certificate only addresses the client side; the server must also require SSL, which is achieved through the parameter group setting.

498
MCQeasy

A developer retrieved a database secret using the AWS CLI as shown. What is the MOST secure way to store and rotate this secret?

A.Store the secret in AWS Secrets Manager and enable automatic rotation with a Lambda function.
B.Store the secret in AWS Systems Manager Parameter Store as a SecureString.
C.Store the secret in a configuration file on the EC2 instance.
D.Use the secret as-is and change it manually every 90 days.
AnswerA

Secrets Manager handles rotation securely.

Why this answer

Secrets Manager can automatically rotate secrets, and the secret should be retrieved using IAM permissions. Option B is insecure. Option C is not best practice.

Option D is not needed.

499
MCQeasy

A company wants to migrate their on-premises Oracle database to Amazon RDS for Oracle. They have a complex data loading process that uses Oracle Data Pump. Which migration approach is MOST efficient and minimizes downtime?

A.Use AWS Schema Conversion Tool (SCT) to convert the schema and then copy data files directly.
B.Use AWS Database Migration Service (DMS) with ongoing replication from the source Oracle database.
C.Take a physical backup of the on-premises database and restore to RDS.
D.Export data using Oracle Data Pump and import into RDS.
AnswerB

DMS supports full load + CDC, reducing downtime.

Why this answer

AWS DMS with ongoing replication (change data capture, CDC) is the most efficient approach for migrating an Oracle database to Amazon RDS for Oracle with minimal downtime. It allows a full load of the existing data followed by continuous replication of changes from the source until cutover, reducing the outage window to seconds or minutes. This directly addresses the requirement to minimize downtime while handling complex data loading processes.

Exam trap

The trap here is that candidates often assume Oracle Data Pump (Option D) is the fastest because it is a native Oracle tool, but they overlook the requirement to minimize downtime, which DMS with CDC addresses by allowing the source to remain operational until the final cutover.

How to eliminate wrong answers

Option A is wrong because AWS SCT is used for schema conversion (e.g., from Oracle to Aurora or PostgreSQL), not for copying data files directly; copying data files is not a supported migration method for RDS for Oracle and would require manual file-level access that RDS does not provide. Option C is wrong because taking a physical backup of an on-premises Oracle database and restoring to RDS is not supported; RDS for Oracle does not allow direct restoration of physical backups from external sources—it requires logical export/import or DMS. Option D is wrong because exporting data using Oracle Data Pump and importing into RDS is a valid method but involves significant downtime as the source database must be quiesced or taken offline during the export and import process, making it less efficient for minimizing downtime compared to DMS with CDC.

500
Multi-Selectmedium

A company is using Amazon DynamoDB with on-demand capacity. The operations team wants to monitor for throttled read requests. Which metric from Amazon CloudWatch should be used to set up an alarm for read throttling?

Select 1 answer
A.ConsumedWriteCapacityUnits
B.ReadThrottleEvents
C.ThrottledPutRecords
D.ProvisionedReadCapacityUnits
E.ThrottledGetRecords
AnswersB

ReadThrottleEvents directly counts throttled read requests on the table.

Why this answer

'ReadThrottleEvents' is a CloudWatch metric that directly indicates throttled read requests on a DynamoDB table. For write throttling, the correct metric is 'WriteThrottleEvents', but it is not among the options given. Therefore, only one option correctly answers the question as revised.

501
MCQmedium

A company stores sensitive customer data in an Amazon S3 bucket. The data is accessed by an Amazon Redshift cluster using the COPY command. The security team wants to ensure that data is encrypted in transit between S3 and Redshift. Which configuration should be used?

A.Use a VPC endpoint for S3 with a bucket policy that denies HTTP.
B.Use the 'SSH' option in the COPY command to encrypt the transfer.
C.Use the 'ENCRYPTED' option with the COPY command and ensure the S3 bucket policy requires HTTPS.
D.Enable S3 server-side encryption on the bucket.
AnswerC

The 'ENCRYPTED' option forces the COPY command to use HTTPS encryption in transit.

Why this answer

The COPY command supports encryption in transit via HTTPS when using the 'ENCRYPTED' option or using an S3 endpoint that enforces HTTPS. Option A is wrong because S3 supports HTTPS, and specifying 'ENCRYPTED' is needed. Option B is wrong because the COPY command does not use SSH; it uses HTTPS.

Option D is wrong because S3 server-side encryption protects data at rest, not in transit.

502
Multi-Selecthard

A company is running an Amazon RDS for SQL Server DB instance with Multi-AZ deployment. The security team wants to ensure that all data at rest is encrypted using a customer-managed key stored in AWS KMS. Which steps must be taken to achieve this? (Choose THREE.)

Select 3 answers
A.Modify the DB instance and enable encryption.
B.Enable Multi-AZ deployment to encrypt data at rest.
C.Copy the snapshot and specify encryption with a KMS key.
D.Create a manual snapshot of the existing DB instance.
E.Restore the DB instance from the encrypted snapshot.
AnswersC, D, E

Copying a snapshot allows you to enable encryption.

Why this answer

To encrypt an existing unencrypted RDS instance, you must create a manual snapshot (D), copy the snapshot with encryption using a KMS key (C), and then restore the DB instance from the encrypted snapshot (E). Option A is incorrect because you cannot enable encryption on an existing DB instance directly. Option B is incorrect because Multi-AZ deployment does not automatically encrypt data.

503
MCQeasy

A company is implementing fine-grained access control for a DynamoDB table named UserSessions. The table has a partition key of 'user_id'. The above IAM policy is attached to an IAM role assumed by the application. What does this policy achieve?

A.Allows the application to perform all operations on the UserSessions table without restrictions
B.Restricts the application to access only items where the partition key matches the user's AWS user ID
C.Allows the application to read but not write items in the UserSessions table
D.Allows the application to access only the UserSessions table but not other tables
AnswerB

The condition uses 'aws:userid' to limit access to items with the corresponding partition key.

Why this answer

The IAM policy uses a condition key `dynamodb:LeadingKeys` with a value of `${aws:userid}`. This restricts access to items in the DynamoDB table where the partition key (`user_id`) matches the unique identifier of the IAM user or role that is making the request. This implements fine-grained access control, ensuring the application can only read or write items belonging to the authenticated user.

Exam trap

The trap here is that candidates often confuse `aws:userid` with the IAM user name or the partition key value, or they assume the policy grants full access (Option A) without noticing the condition that enforces row-level security.

How to eliminate wrong answers

Option A is wrong because the policy explicitly restricts access based on the partition key, so it does not allow all operations without restrictions. Option C is wrong because the policy does not specify any `Action` or `Effect` that limits operations to read-only; it allows all DynamoDB actions on the table, subject to the condition. Option D is wrong because the policy's `Resource` element is scoped to the `UserSessions` table ARN, but the condition is what restricts access within that table, not the ability to access other tables (which would be denied by default if not explicitly allowed).

504
MCQmedium

A financial services company is migrating an on-premises Oracle database to Amazon RDS for Oracle. The database has a high volume of write transactions and requires minimal downtime during migration. Which AWS service or feature should be used to replicate data continuously to the target RDS instance during the migration?

A.Amazon RDS Read Replica
B.Amazon RDS Multi-AZ deployment
C.AWS Database Migration Service (AWS DMS) with ongoing replication
D.AWS Schema Conversion Tool (AWS SCT)
AnswerC

AWS DMS can perform full load and then continuously replicate changes via CDC.

Why this answer

AWS Database Migration Service (AWS DMS) with ongoing replication (change data capture, CDC) is the correct choice because it continuously captures and applies changes from the source Oracle database to the target Amazon RDS for Oracle instance, enabling near-zero downtime during migration. This is achieved by using Oracle's redo logs to stream transactions in real time, which meets the high write volume and minimal downtime requirements.

Exam trap

The trap here is that candidates confuse continuous replication with high-availability features like Multi-AZ or read replicas, but those services cannot ingest data from an on-premises source; only AWS DMS with CDC provides the necessary ongoing replication for a live migration with minimal downtime.

How to eliminate wrong answers

Option A is wrong because Amazon RDS Read Replica is designed for read scaling and asynchronous replication from an RDS source, not for migrating an on-premises Oracle database; it cannot connect to an external source. Option B is wrong because Amazon RDS Multi-AZ deployment provides high availability by synchronously replicating data to a standby instance in another Availability Zone, but it does not support continuous replication from an on-premises database. Option D is wrong because the AWS Schema Conversion Tool (AWS SCT) is used to convert database schemas and code for heterogeneous migrations, not for continuous data replication.

505
MCQmedium

A company is deploying a new multi-AZ Aurora MySQL database. The application requires read-heavy workloads and low latency. Which configuration will best meet these requirements?

A.Enable Multi-AZ and use the secondary for read traffic.
B.Deploy a single instance with one read replica in the same AZ.
C.Enable Aurora Auto Scaling with a target metric of average CPU utilization.
D.Use an RDS Proxy in front of the database.
AnswerC

Aurora Auto Scaling adds reader instances automatically to handle read-heavy workloads, improving latency.

Why this answer

Aurora Auto Scaling dynamically adds reader instances based on the average CPU utilization metric, which directly addresses the read-heavy workload requirement by distributing read traffic across multiple replicas. This configuration ensures low latency by scaling out read capacity automatically as demand increases, without manual intervention.

Exam trap

The trap here is that candidates confuse Multi-AZ with read scaling, assuming the standby instance can serve reads, but in Aurora Multi-AZ, the secondary is a writer instance for failover only, not a reader.

How to eliminate wrong answers

Option A is wrong because in Aurora Multi-AZ, the secondary (writer) instance is not used for read traffic; it only provides failover support, and using it for reads would degrade write performance and violate Aurora's architecture. Option B is wrong because deploying a single instance with one read replica in the same AZ does not provide high availability or fault tolerance, and a single replica cannot handle heavy read loads effectively, leading to potential latency issues. Option D is wrong because RDS Proxy manages connection pooling and reduces database load from connection churn, but it does not directly address read-heavy workloads or scale read capacity; it is a connection management tool, not a read scaling solution.

506
MCQmedium

A company is deploying a new web application on AWS. The application uses Amazon RDS for MySQL as its database. The database must be highly available and automatically failover in the event of an AZ outage. The company also needs to offload read traffic from the primary database to improve performance. The application read-to-write ratio is 80:20. The database workload is variable, with occasional spikes. The company wants a cost-effective solution that scales read capacity automatically. The operations team has limited experience with AWS. Which solution should the company implement?

A.Deploy an RDS for MySQL instance with Multi-AZ and one or more read replicas, and use Application Auto Scaling to add replicas based on CPU utilization
B.Deploy an RDS for MySQL instance with Multi-AZ and enable automatic scaling of the instance size
C.Deploy an Amazon Aurora MySQL cluster with Multi-AZ and enable Aurora Auto Scaling for read replicas
D.Deploy an RDS for MySQL instance with multiple read replicas and use a custom script to promote a read replica in case of failure
AnswerC

Aurora provides automatic failover and auto-scaling of read replicas, and is MySQL-compatible, making it a cost-effective and simpler solution.

Why this answer

This solution is correct because Amazon Aurora is MySQL-compatible, provides high availability and automatic failover through its Multi-AZ deployment, and includes Aurora Auto Scaling which automatically adjusts the number of read replicas based on workload, scaling read capacity automatically. This is cost-effective as you pay only for the replicas you use and it requires minimal management, making it ideal for teams with limited AWS experience. Option A is incorrect because read replicas do not auto-scale automatically; Application Auto Scaling for replicas requires custom setup.

Option B is incorrect because scaling instance size does not offload read traffic or provide read scaling. Option D is incorrect because it lacks automatic failover and requires manual intervention.

507
MCQmedium

A retail company uses Amazon RDS for PostgreSQL as the backend for its e-commerce platform. During a flash sale, the database experienced high CPU utilization and increased the number of active connections. The application team reported that some queries timed out. The database specialist reviewed the slow query log and found that several queries were performing sequential scans on large tables due to missing indexes. The specialist created the necessary indexes, but the issue persists for some queries. Upon further investigation, the specialist notices that the query planner is still choosing sequential scans for some queries. What should the database specialist do to ensure the query planner uses the indexes?

A.Increase maintenance_work_mem to speed up index creation.
B.Decrease the random_page_cost to make indexes more attractive.
C.Run the ANALYZE command to update table statistics.
D.Set enable_seqscan to off to force index usage.
AnswerC

Updated statistics help the planner use indexes.

Why this answer

Running ANALYZE updates the table statistics used by the query planner, allowing it to make informed decisions about index usage. Without updated statistics, the planner may still choose sequential scans even after indexes are created. Option A is incorrect because increasing maintenance_work_mem speeds up index creation but does not affect query planning.

Option B is incorrect because decreasing random_page_cost might make indexes more attractive, but it is a global setting that could have unintended consequences and does not address the root cause of stale statistics. Option D is incorrect because disabling sequential scans (enable_seqscan=off) forces index usage but can lead to suboptimal plans, especially if the index is not the most efficient access method.

508
Multi-Selectmedium

A company is migrating a 200 GB Oracle database to Amazon Aurora MySQL. They want to minimize downtime and ensure data consistency. Which two services should they use together? (Choose TWO.)

Select 2 answers
A.AWS Snowball Edge
B.AWS Schema Conversion Tool (SCT)
C.AWS Database Migration Service (DMS)
D.AWS Data Pipeline
E.Amazon EC2 with Oracle installed
AnswersB, C

SCT converts Oracle schema to Aurora MySQL.

Why this answer

AWS Schema Conversion Tool (SCT) is required to convert the Oracle database schema (including stored procedures, functions, and data types) to a format compatible with Amazon Aurora MySQL. AWS Database Migration Service (DMS) then performs the continuous data replication from Oracle to Aurora MySQL, enabling near-zero downtime migration while maintaining data consistency through Change Data Capture (CDC).

Exam trap

The trap here is that candidates often assume AWS DMS alone can handle heterogeneous migrations, forgetting that SCT is mandatory for schema conversion when moving from Oracle to Aurora MySQL, as DMS only handles data movement and cannot convert incompatible database objects.

509
MCQeasy

A financial services company uses Amazon DynamoDB to store transaction records. Each transaction has a unique transaction_id as the partition key and a timestamp as the sort key. The application frequently queries all transactions for a given customer within a date range. However, customer_id is not an attribute indexed for querying. The company wants to optimize these queries without redesigning the entire table schema. Which action should the company take?

A.Change the table's partition key to customer_id and use a composite sort key.
B.Create a Local Secondary Index (LSI) on customer_id.
C.Create a Global Secondary Index (GSI) with customer_id as the partition key and timestamp as the sort key.
D.Use the Scan operation with a filter expression for customer_id and timestamp.
AnswerC

A GSI allows querying by customer_id and timestamp range without modifying the base table.

Why this answer

Creating a Global Secondary Index (GSI) with customer_id as the partition key and timestamp as the sort key allows efficient querying of all transactions for a given customer within a date range without redesigning the base table. The GSI provides a new access pattern with its own partition and sort keys, enabling the Query operation on customer_id and timestamp, which is far more efficient than a Scan. This approach preserves the existing table schema and supports the required query pattern with minimal overhead.

Exam trap

The trap here is that candidates often confuse Local Secondary Indexes (LSIs) with Global Secondary Indexes (GSIs), assuming an LSI can be added later or can use a different partition key, when in fact LSIs must share the base table's partition key and can only be created at table creation time.

How to eliminate wrong answers

Option A is wrong because changing the table's partition key to customer_id would require a full table redesign, data migration, and application downtime, which contradicts the requirement to avoid redesigning the entire table schema. Option B is wrong because a Local Secondary Index (LSI) can only be created at table creation time and must use the same partition key as the base table (transaction_id), so it cannot index on customer_id as a partition key for range queries. Option D is wrong because using the Scan operation with a filter expression for customer_id and timestamp is inefficient, as it reads every item in the table and then filters, incurring high read capacity consumption and latency, especially for large tables.

510
MCQhard

A company is deploying a new web application using Amazon RDS for PostgreSQL. The application requires read-heavy workloads and automatic failover. Which configuration should be used?

A.Multi-AZ deployment without Read Replicas
B.Single-AZ with a Read Replica in the same region
C.Multiple Read Replicas in different regions
D.Multi-AZ deployment with one or more Read Replicas
AnswerD

Provides HA and read scaling.

Why this answer

A Multi-AZ deployment provides automatic failover to a standby instance in a different Availability Zone, ensuring high availability. Adding one or more Read Replicas offloads read-heavy workloads from the primary instance, improving performance. This combination meets both the read-heavy and automatic failover requirements for the application.

Exam trap

The trap here is that candidates often assume Multi-AZ alone handles read scaling, but it does not; the standby in Multi-AZ is not accessible for reads, so Read Replicas are required for read-heavy workloads.

How to eliminate wrong answers

Option A is wrong because while Multi-AZ provides automatic failover, it does not include Read Replicas, so read-heavy workloads would still hit the primary instance, causing performance bottlenecks. Option B is wrong because Single-AZ with a Read Replica lacks automatic failover; if the primary fails, the Read Replica must be manually promoted, causing downtime. Option C is wrong because multiple Read Replicas in different regions address read scaling but do not provide automatic failover; failover requires a Multi-AZ deployment or manual promotion.

511
MCQmedium

A company has a document database workload on Amazon DynamoDB that stores user session data. The application frequently updates session attributes (e.g., last activity timestamp). The current design stores the entire session as a single item and updates the entire item on each session activity. This is causing high write costs and throttling. Which design pattern would reduce write costs and improve performance?

A.Increase the write capacity units (WCUs) on the table.
B.Use UpdateItem with an update expression to modify only the changed attributes.
C.Implement DynamoDB Accelerator (DAX) to cache the session data.
D.Split the session item into multiple items, one per attribute.
AnswerB

Update expressions only write the changed attributes, consuming fewer write capacity units.

Why this answer

Using UpdateItem with an update expression allows you to modify only the specific attributes that changed (e.g., last activity timestamp) instead of rewriting the entire item. This reduces write consumption to a fraction of the original cost, since DynamoDB charges based on the size of the written data, and partial updates write only the changed attribute bytes. This directly addresses the high write costs and throttling caused by full-item overwrites.

Exam trap

The trap here is that candidates often confuse scaling solutions (increasing WCUs or adding DAX) with optimization patterns, failing to recognize that the real issue is the write amplification caused by full-item updates rather than insufficient capacity or read performance.

How to eliminate wrong answers

Option A is wrong because increasing write capacity units (WCUs) only raises the throughput limit but does not reduce the cost per write; it would increase costs further and does not solve the root cause of writing the entire item. Option C is wrong because DynamoDB Accelerator (DAX) is an in-memory cache that improves read performance, not write cost or write throttling; it does not reduce the amount of data written per update. Option D is wrong because splitting a session item into multiple items per attribute would require multiple write operations for each session update, increasing write costs and complexity, and DynamoDB charges per write request regardless of item size.

512
MCQeasy

A company is migrating an on-premises Oracle OLTP workload to AWS. The database has complex stored procedures and requires minimal code changes. Which AWS database service is the most suitable target?

A.Amazon Redshift
B.Amazon DynamoDB
C.Amazon Aurora PostgreSQL
D.Amazon RDS for Oracle
AnswerD

Minimal code changes required.

Why this answer

Amazon RDS for Oracle is the most suitable target because it provides native Oracle compatibility, allowing the existing complex stored procedures and PL/SQL code to run with minimal or no changes. This minimizes migration risk and effort, which is the primary requirement for an OLTP workload with complex stored procedures.

Exam trap

The trap here is that candidates may assume Aurora PostgreSQL is a drop-in replacement for Oracle due to its PostgreSQL compatibility features, but it still requires significant code changes for complex stored procedures, whereas RDS for Oracle avoids this entirely.

How to eliminate wrong answers

Option A is wrong because Amazon Redshift is a columnar data warehouse optimized for analytical queries, not OLTP workloads, and it does not support Oracle stored procedures or PL/SQL. Option B is wrong because Amazon DynamoDB is a NoSQL key-value and document database that does not support relational features like stored procedures, joins, or complex transactions required by the Oracle OLTP workload. Option C is wrong because Amazon Aurora PostgreSQL uses PostgreSQL syntax and PL/pgSQL, which would require significant code changes to migrate complex Oracle stored procedures and PL/SQL logic.

513
MCQeasy

A company runs a MySQL database on Amazon RDS for an e-commerce platform. The application performs frequent INSERT and UPDATE operations on the 'orders' table. The team notices an increase in disk I/O and CPU usage. They want to optimize the database for write-heavy workloads without changing the application. Which option is the MOST effective?

A.Enable Multi-AZ deployment for redundancy
B.Change the storage engine to MyISAM
C.Increase the InnoDB buffer pool size
D.Upgrade to Amazon Aurora MySQL
AnswerC

A larger buffer pool reduces disk I/O by caching data and indexes in memory.

Why this answer

Increasing the InnoDB buffer pool size allows more data and indexes to be cached in memory, reducing disk I/O for write operations by delaying writes and enabling more efficient page merging. This directly addresses the high disk I/O and CPU usage from frequent INSERT and UPDATE operations without requiring application changes.

Exam trap

The DBS-C01 exam often tests the misconception that Multi-AZ improves performance, when in fact it only provides redundancy and can slightly increase write latency due to synchronous replication to the standby instance.

How to eliminate wrong answers

Option A is wrong because Multi-AZ deployment provides high availability and automatic failover, but does not optimize write performance or reduce disk I/O. Option B is wrong because MyISAM does not support transactions or row-level locking, and it uses table-level locking which would severely degrade concurrent write performance in a write-heavy workload. Option D is wrong because upgrading to Aurora MySQL would require application changes (different endpoint, potential compatibility issues) and is not the most effective immediate optimization; increasing the buffer pool size is a simpler, non-disruptive change.

514
MCQeasy

A startup is building a multi-tenant SaaS application where each tenant's data must be isolated. The data model is relational with complex joins. Which database deployment model is most appropriate?

A.Use a single Amazon DynamoDB table with a tenant_id partition key
B.Use a single Amazon Redshift cluster with tenant_id distribution key
C.Provision a separate Amazon RDS instance for each tenant
D.Use a single Amazon RDS database with a tenant_id column on every table
AnswerC

Separate instances ensure complete data isolation and independent scaling.

Why this answer

A multi-tenant SaaS application requiring strict data isolation with complex relational joins demands separate databases per tenant. Amazon RDS provides full relational capabilities (ACID transactions, complex joins) and provisioning separate RDS instances ensures complete tenant isolation, preventing cross-tenant data leakage and allowing independent backup, scaling, and performance tuning for each tenant.

Exam trap

The trap here is that candidates often choose logical isolation (Option D) thinking it is sufficient, but the exam emphasizes that strict data isolation for multi-tenant SaaS with complex relational joins requires physical database separation to prevent cross-tenant data leaks and ensure compliance.

How to eliminate wrong answers

Option A is wrong because Amazon DynamoDB is a NoSQL key-value/document database that does not support complex joins or relational queries; using a single table with tenant_id partition key would require application-level joins and cannot enforce relational integrity. Option B is wrong because Amazon Redshift is a columnar data warehouse optimized for analytical queries on large datasets, not for transactional OLTP workloads with complex joins per tenant; it lacks row-level isolation and is not designed for multi-tenant SaaS with per-tenant data isolation. Option D is wrong because a single RDS database with a tenant_id column on every table provides logical isolation only, not physical isolation; a bug in a query or a missing WHERE clause could expose one tenant's data to another, violating the strict isolation requirement.

515
MCQhard

A company is using Amazon DynamoDB Accelerator (DAX) for caching. The security team is concerned about data in transit between the application and DAX. What should the team do to ensure that all traffic to DAX is encrypted?

A.Launch the DAX cluster in a private subnet with a VPC endpoint.
B.Enable encryption in transit when creating the DAX cluster.
C.Use AWS Certificate Manager to issue a certificate for the DAX cluster.
D.Use client-side encryption to encrypt data before sending it to DAX.
AnswerB

DAX supports TLS encryption in transit when enabled at cluster creation.

Why this answer

DAX supports encryption in transit, which must be enabled when creating the cluster. Option A is incorrect because launching in a private subnet with a VPC endpoint does not encrypt traffic; it only provides private connectivity. Option C is incorrect because DAX manages its own encryption certificates and does not use AWS Certificate Manager.

Option D is incorrect because client-side encryption encrypts data at the application layer, but it does not ensure encryption in transit between the application and DAX; enabling encryption in transit on the DAX cluster is required.

516
Multi-Selecteasy

Which TWO are valid use cases for Amazon ElastiCache for Redis? (Choose 2)

Select 2 answers
A.Storing graph data with relationships
B.Session management for web applications
C.Running complex analytical queries on large datasets
D.Caching frequently accessed database queries to reduce load on RDS
E.Persistent storage of relational data
AnswersB, D

Redis is often used for session storage due to low latency.

Why this answer

Amazon ElastiCache for Redis is an in-memory data store ideal for session management because it provides sub-millisecond latency for storing and retrieving session tokens, supports TTL-based key expiration to automatically clean up stale sessions, and offers atomic operations like SETEX for safe session creation. This makes it a perfect fit for stateless web applications that need to offload session state from the application server.

Exam trap

The trap here is that candidates often confuse caching (option D) with persistent storage (option E) or assume that Redis's data structures (like sorted sets) can handle graph relationships (option A), but Redis lacks the graph traversal and indexing capabilities of a dedicated graph database.

517
Multi-Selecthard

A company is deploying a new Amazon RDS for Oracle database in a VPC. The database must be accessed by an application running on an EC2 instance in a different subnet. Which THREE steps are required to allow this access?

Select 3 answers
A.Create a VPC peering connection between the subnets.
B.Attach an Internet Gateway to the VPC.
C.Configure the network ACL for the RDS subnet to allow inbound traffic from the EC2 subnet.
D.Ensure the VPC has the 'Enable DNS hostnames' attribute set to true.
E.Add an inbound rule to the RDS security group that allows traffic from the EC2 security group.
AnswersC, D, E

NACLs provide stateless filtering.

Why this answer

The security group of the RDS instance must allow inbound traffic from the EC2 security group, which is option E. The VPC must have the 'Enable DNS hostnames' attribute set to true (option D) so that the RDS endpoint resolves correctly. Additionally, the network ACL for the RDS subnet must allow inbound traffic from the EC2 subnet (option C).

Option A (VPC peering) is not needed because both resources are in the same VPC. Option B (Internet Gateway) is not needed because traffic is within the VPC and does not require internet access.

518
Multi-Selectmedium

A company needs to migrate a 1 TB Oracle database to Amazon RDS for Oracle. The migration must have minimal downtime and the source database is running on-premises. Which TWO AWS services should be used together to achieve this?

Select 2 answers
A.AWS Database Migration Service (DMS)
B.AWS Schema Conversion Tool (SCT)
C.AWS CloudEndure Migration
D.AWS Snowball
E.AWS Direct Connect
AnswersA, D

AWS DMS is correct. It handles the full load and CDC replication for minimal downtime.

Why this answer

To migrate a 1 TB Oracle database to Amazon RDS for Oracle with minimal downtime, two services are required. AWS Database Migration Service (DMS) performs the initial full load and then uses Change Data Capture (CDC) to continuously replicate ongoing changes, minimizing downtime. AWS Direct Connect establishes a dedicated network connection between the on-premises environment and AWS, ensuring stable and high-bandwidth data transfer necessary for the large database migration.

Without Direct Connect, the replication could be impacted by internet latency or outages.

Exam trap

The trap is that candidates may assume AWS DMS alone is sufficient for a homogeneous migration. However, for a 1 TB database with minimal downtime, a dedicated network service like AWS Direct Connect is essential to maintain consistent throughput. Candidates might incorrectly choose AWS SCT, which is only needed for heterogeneous migrations, or AWS Snowball, which is for offline transfer and not suitable for minimal downtime.

519
MCQeasy

A company wants to deploy a DynamoDB table that requires consistent single-digit millisecond latency regardless of traffic spikes. Which DynamoDB capacity mode should be selected?

A.DynamoDB Accelerator (DAX) enabled.
B.Provisioned capacity mode with auto scaling.
C.On-demand capacity mode.
D.Provisioned capacity mode with fixed read/write capacity.
AnswerC

Automatically scales to handle any traffic level.

Why this answer

On-demand capacity mode is the correct choice because it automatically scales read and write capacity up and down based on actual traffic, eliminating the need for capacity planning and ensuring consistent single-digit millisecond latency even during unpredictable traffic spikes. This mode is designed for workloads with variable or bursty traffic patterns where latency sensitivity is critical.

Exam trap

The DBS-C01 exam often tests the misconception that DAX is a capacity mode or that auto scaling provides instant burst protection, when in fact only on-demand mode guarantees zero throttling during sudden traffic spikes without pre-provisioning.

How to eliminate wrong answers

Option A is wrong because DynamoDB Accelerator (DAX) is an in-memory caching service that reduces read latency but does not change the capacity mode; it is an add-on, not a capacity mode itself. Option B is wrong because provisioned capacity with auto scaling adjusts capacity based on utilization metrics, but it cannot react instantly to sudden traffic spikes, potentially causing throttling and latency degradation during rapid bursts. Option D is wrong because fixed provisioned capacity requires manual capacity planning and cannot handle traffic spikes without risking throttling or excessive cost, leading to inconsistent latency.

520
MCQhard

A company is running a MongoDB-compatible Amazon DocumentDB cluster. The application experiences high write latency during peak hours. The database administrator checks the CloudWatch metrics and notices that the Write IOPS metric is consistently at the maximum for the instance size. What should the administrator do to reduce write latency?

A.Increase the instance size to a larger instance class with higher IOPS limits.
B.Switch to a memory-optimized instance class.
C.Increase the allocated storage size to improve I/O performance.
D.Enable Multi-AZ deployment to offload writes to the standby.
AnswerA

Increasing the instance size provides more IOPS capacity, which can reduce write latency if the instance is hitting IOPS limits.

Why this answer

Increasing the instance size provides more IOPS capacity, which can reduce write latency if the instance is hitting IOPS limits. Option B is incorrect because switching to a memory-optimized instance class does not directly increase IOPS capacity; it improves memory performance, not I/O throughput. Option C is incorrect because increasing allocated storage size may increase IOPS if using gp2, but the issue is hitting the maximum IOPS for the current instance, and simply increasing storage does not guarantee a proportional increase in IOPS limits.

Option D is incorrect because enabling Multi-AZ adds a standby replica for high availability but does not offload writes or increase write IOPS capacity on the primary.

521
MCQeasy

A company is running a production Amazon DynamoDB table and notices that read requests are being throttled. The table has on-demand capacity mode enabled. Which action should the database specialist take to troubleshoot the throttling?

A.Check the CloudWatch metric 'ThrottledRequests' for the table and review 'SystemErrors' to identify hot partitions.
B.Enable auto scaling on the table to automatically adjust capacity.
C.Enable DynamoDB Accelerator (DAX) to reduce read load on the table.
D.Switch the table to provisioned capacity mode and increase the read capacity units.
AnswerA

Throttling with on-demand can be due to a hot partition; CloudWatch metrics help identify it.

Why this answer

With on-demand capacity, throttling often results from hot partitions. The CloudWatch metric 'ThrottledRequests' helps detect throttling, and reviewing 'SystemErrors' can indicate partition-level errors, aiding in identifying hot partitions. Option B is incorrect because auto scaling is only supported for provisioned capacity mode, not on-demand.

Option C is incorrect because enabling DAX can reduce read load but does not address the root cause of throttling due to hot partitions; it is not a troubleshooting step. Option D is incorrect because switching to provisioned capacity and increasing RCUs is a remediation action, not a troubleshooting action.

522
MCQmedium

Refer to the exhibit. An IAM policy is attached to an IAM user that performs database migrations. When the user tries to start an AWS DMS replication task that writes to the RDS instance 'targetdb', the task fails with an access denied error. Which additional permission is required?

A.dms:CreateEndpoint
B.dms:CreateReplicationInstance
C.dms:DescribeReplicationInstances
D.rds:ModifyDBInstance on the source database
AnswerC

The user may need to describe replication instances to select one for the task.

Why this answer

The DMS replication task requires permission to describe the replication instance that is being used to run the task. The `dms:DescribeReplicationInstances` action allows the IAM user to retrieve metadata about the replication instance, which DMS needs to validate the instance state and configuration before starting the task. Without this permission, the start task operation fails with an access denied error even if the user has permissions to start the task itself.

Exam trap

The trap here is that candidates assume the error is due to missing permissions for the target database (RDS) or for creating resources, when in fact DMS requires read-only describe permissions on the replication instance to validate its state before starting a task.

How to eliminate wrong answers

Option A is wrong because `dms:CreateEndpoint` is used to create source or target endpoints, not to start an existing replication task; the error occurs during task start, not endpoint creation. Option B is wrong because `dms:CreateReplicationInstance` is required to provision a new replication instance, but the task failure is about starting a task on an existing instance, not creating one. Option D is wrong because `rds:ModifyDBInstance` on the source database is unrelated to starting a DMS task; the error is from DMS, not from RDS, and modifying the source database is not a prerequisite for task execution.

523
MCQhard

A company is using Amazon DynamoDB with global tables. The application team reports that data written in one region is not immediately available in another region. The database specialist needs to monitor the replication lag. Which CloudWatch metric should be used?

A.Monitor the 'ConsumedWriteCapacityUnits' metric in both regions.
B.Monitor the 'PendingReplicationCount' metric in the source region.
C.Monitor the 'ThrottledRequests' metric in the source region.
D.Monitor the 'ReplicationLatency' metric in the replica region.
AnswerD

This metric directly measures the replication lag between regions.

Why this answer

'ReplicationLatency' is the CloudWatch metric that measures the time between an update to a DynamoDB global table in the source region and its appearance in the replica region. Option A is wrong because 'ConsumedWriteCapacityUnits' measures the amount of write capacity consumed, not replication lag. Option B is wrong because 'PendingReplicationCount' shows the number of items waiting to be replicated, not the time delay.

Option C is wrong because 'ThrottledRequests' indicates that requests are being throttled, which is unrelated to replication lag.

524
MCQeasy

A database administrator notices that an Amazon RDS for MySQL instance's storage is filling up unexpectedly. The administrator has enabled automated backups and retains them for 7 days. Which of the following actions would help reduce storage consumption without losing the ability to perform point-in-time recovery?

A.Modify the DB instance to a smaller instance class
B.Reduce the backup retention period to 1 day
C.Delete manual DB snapshots
D.Disable automated backups
AnswerB

Reducing backup retention minimizes the volume of automated backup data stored, which directly affects the storage consumption associated with backups. Point-in-time recovery is still possible for the retained period.

Why this answer

Reducing the backup retention period reduces the amount of storage consumed by automated backup data, while still allowing point-in-time recovery for the duration of the retention period. Option A is incorrect because changing the instance class does not affect storage consumption. Option C is incorrect because deleting manual snapshots does not reduce the automated backup storage that is likely causing the unexpected filling, and manual snapshots are not necessary for point-in-time recovery.

Option D is incorrect because disabling automated backups eliminates point-in-time recovery capability.

525
MCQmedium

A company is migrating a 2 TB Oracle database to Amazon Aurora PostgreSQL. They need to minimize downtime and ensure data consistency. Which migration strategy should they use?

A.Use AWS DMS with full load only, then cut over during a maintenance window.
B.Use AWS DMS with ongoing replication (change data capture) from Oracle to Aurora.
C.Export the database to Amazon S3 using Oracle Data Pump, then import into Aurora.
D.Use pg_dump to export the database and pg_restore to import into Aurora.
AnswerB

DMS CDC captures ongoing changes, allowing near-zero downtime migration.

Why this answer

AWS DMS with ongoing replication (change data capture) is the correct strategy because it allows continuous synchronization of changes from the source Oracle database to the target Aurora PostgreSQL, minimizing downtime to just the final cutover window. This approach ensures data consistency by applying transactional changes in near real-time, which is essential for a 2 TB database where a full load alone would require a lengthy maintenance window and risk data divergence.

Exam trap

The trap here is that candidates often assume a full-load-only approach (Option A) is sufficient for large databases, underestimating the downtime required to stop writes and the risk of data inconsistency, while overlooking that DMS's ongoing replication is the only option that combines minimal downtime with continuous data synchronization.

How to eliminate wrong answers

Option A is wrong because full load only captures a point-in-time snapshot; any changes made after the load starts are lost, requiring a long maintenance window to stop all writes, which contradicts the goal of minimizing downtime. Option C is wrong because exporting to Amazon S3 using Oracle Data Pump and then importing into Aurora is a manual, offline process that requires the source database to be quiesced or taken offline, causing significant downtime and lacking built-in change data capture for ongoing sync. Option D is wrong because pg_dump and pg_restore are PostgreSQL-native tools that cannot connect to or extract data from an Oracle source database; they are designed for PostgreSQL-to-PostgreSQL migrations only.

Page 6

Page 7 of 23

Page 8