Troubleshooting and Optimization

154. An AWS Lambda function returns HTTP 502 through Amazon API Gateway. Which cause is most likely?

1. A Increase the function's memory allocation because the invocation is running out of memory.
2. B Correct the function to return a response object containing statusCode and body for the proxy integration.
3. C Reduce the function's reserved concurrency because invocations are being throttled.
4. D Deploy the API stage because the current configuration has not been deployed.

155. An application receives ThrottlingException responses from an AWS service during traffic spikes. Which solution meets these requirements?

1. A Fail the request immediately and raise an alert to an operator.
2. B Open additional concurrent connections so more requests are in flight.
3. C Retry the failed calls immediately in a loop until they succeed.
4. D Retry the failed calls with exponential backoff and jitter up to a bounded number of attempts.

156. An AWS Lambda function shows duration close to its limit while memory usage stays well below the configured allocation. Which step should the developer take first?

1. A Reduce the function timeout so failing invocations end sooner.
2. B Determine whether the function is waiting on network or storage rather than consuming CPU, because raising memory only adds CPU.
3. C Enable provisioned concurrency so execution environments are initialized in advance.
4. D Increase the memory allocation further so the function receives more CPU.

157. An application's Amazon DynamoDB writes intermittently fail with a validation error stating that the item size exceeds the limit. Which solution meets these requirements?

1. A Enable DynamoDB Streams so oversized items are captured for reprocessing.
2. B Split the write across a BatchWriteItem call containing the same item.
3. C Increase the table's provisioned write capacity units to accommodate the larger items.
4. D Store the large payload in Amazon S3 and keep only a pointer to the object in the DynamoDB item.

158. A developer must search several gigabytes of AWS Lambda logs for one correlation ID across a two-day window. Which solution meets these requirements?

1. A Run a CloudWatch Logs Insights query that filters on the correlation ID across the log groups and time range.
2. B Download the log streams to a workstation and search them locally.
3. C Increase the log group retention period and repeat the search.
4. D Enable AWS X-Ray active tracing for the period already elapsed.

159. An Amazon DynamoDB table shows occasional throttling on one partition while overall provisioned capacity is not exhausted. Which solution provides the durable fix?

1. A Increase the table's provisioned read and write capacity units.
2. B Enable DynamoDB auto scaling on the table's provisioned capacity.
3. C Create a local secondary index on the frequently accessed attribute.
4. D Redesign the partition key to a higher-cardinality value so traffic is distributed across more partitions.

160. An application writes unstructured text to logs, making incident investigation slow. Which solution meets these requirements?

1. A Write the logs to Amazon S3 instead of CloudWatch Logs.
2. B Emit structured JSON log entries with consistent field names so queries can filter on fields directly.
3. C Increase the CloudWatch Logs retention period so more history is available.
4. D Reduce the application's log level so only errors are recorded.

161. An AWS Lambda function intermittently receives server errors from a downstream service. Which client behaviour should the developer implement?

1. A Return a success response to the caller and log the failure for later reconciliation.
2. B Retry with exponential backoff and jitter for a bounded number of attempts, then fail with a clear error.
3. C Fail the invocation immediately without retrying.
4. D Retry indefinitely until the downstream call succeeds.

162. An Amazon API Gateway endpoint returns HTTP 429 to callers during peak traffic. Which cause is most likely?

1. A Deploy the API stage, because the current configuration has not been deployed.
2. B Increase the integration timeout, because the backend exceeded the response deadline.
3. C Raise the throttling limits on the stage, method, or usage plan, because a configured limit has been reached.
4. D Correct the backend response format, because the integration returned a malformed payload.

163. A developer must identify which AWS Lambda invocations exceeded three seconds over the past week across millions of log events. Which solution meets these requirements?

1. A Enable AWS X-Ray active tracing and review the traces for the past week.
2. B Export the log groups to Amazon S3 and search the objects with Amazon Athena.
3. C Run a CloudWatch Logs Insights query that parses the duration from the report lines and filters on it.
4. D Increase the log group retention period and review the logs manually.

164. A developer must find which downstream call causes intermittent latency in a request that traverses Amazon API Gateway, two AWS Lambda functions, and Amazon DynamoDB. Which solution meets these requirements?

1. A Enable VPC Flow Logs on the subnets hosting the functions.
2. B Query AWS CloudTrail event history for the affected period.
3. C Compare CloudWatch metrics for each service over the affected period.
4. D Enable AWS X-Ray active tracing and examine the service map and trace segments.

165. A developer must record a value in an AWS X-Ray trace so that traces can later be filtered on it. Which solution meets these requirements?

1. A Record the value in the subsegment name so it appears in the service map.
2. B Record the value as metadata, which X-Ray stores with the trace.
3. C Record the value as an annotation, which X-Ray indexes for filter expressions.
4. D Record the value in a sampling rule applied to the service.

166. An AWS Lambda function must publish a business metric without adding a synchronous API call to each invocation. Which solution meets these requirements?

1. A Write the metric as a structured log entry in the CloudWatch embedded metric format.
2. B Create a CloudWatch alarm on the function's Invocations metric.
3. C Add the metric value as an X-Ray metadata field on each trace.
4. D Call PutMetricData at the end of each invocation.

167. An operations team must be notified when an application approaches its AWS Lambda concurrency quota. Which solution meets these requirements?

1. A Increase the account concurrency quota so the limit is never approached.
2. B Create a CloudWatch alarm on the ConcurrentExecutions metric at a percentage of the quota and publish to an Amazon SNS topic.
3. C Create a CloudWatch alarm on the function's Errors metric and publish to an Amazon SNS topic.
4. D Review the Service Quotas console for the account each week.

168. A distributed application must record which customer tier each traced request belongs to so that traces can be filtered by tier. Which solution meets these requirements?

1. A Add the customer tier to a CloudWatch custom metric dimension.
2. B Add the customer tier to the segment name so it appears in the service map.
3. C Add the customer tier as AWS X-Ray metadata on the segment so it is stored with the trace.
4. D Add the customer tier as an AWS X-Ray annotation on the segment so it is indexed for filter expressions.

169. An application writes log entries that must record the request identifier, the operation, and the outcome in a form that can be queried without regular expressions. Which solution meets these requirements?

1. A Emit one log entry per value so each can be searched independently.
2. B Emit each log entry as a JSON object with consistent field names for the identifier, operation, and outcome.
3. C Emit the log entries at debug level so more detail is captured.
4. D Emit each log entry as a formatted sentence containing the same three values.

170. An AWS Lambda function connecting to Amazon RDS exhausts the database connection limit under load. Which solution meets these requirements?

1. A Change the database to a larger instance class with a higher connection limit.
2. B Place Amazon RDS Proxy between the function and the database so connections are pooled and reused.
3. C Open a new database connection inside the handler for every invocation.
4. D Increase the function's memory allocation so invocations complete more quickly.

171. An application makes many small sequential calls to Amazon DynamoDB, and latency is dominated by connection setup. Which solution meets these requirements?

1. A Reduce the SDK's socket timeout so failed connections are retried sooner.
2. B Increase the table's provisioned read capacity units.
3. C Create a new SDK client instance for each request to avoid shared state.
4. D Configure the SDK to reuse connections with HTTP keep-alive enabled.

172. A team must determine the AWS Lambda memory setting that minimizes cost per invocation for a CPU-bound function. Which solution meets these requirements?

1. A Configure the function with the minimum memory setting to minimize the per-millisecond price.
2. B Measure duration and cost across a range of memory settings and select the setting with the lowest cost per invocation.
3. C Configure the function's memory to match the size of its deployment package.
4. D Configure the function with the maximum memory setting to minimize the duration.

173. An application's Amazon DynamoDB reads intermittently fail with ProvisionedThroughputExceededException during brief spikes. Which combination of steps meets these requirements? (Select TWO.)

1. A Retry the failed reads immediately in a tight loop until they succeed.
2. B Enable auto scaling on the table or convert it to on-demand capacity mode.
3. C Increase the size of each item so that fewer read requests are needed.
4. D Implement exponential backoff with jitter in the client's retry logic.
5. E Disable the SDK's built-in retry behaviour so failures surface immediately.

174. A CloudWatch Logs subscription filter forwards every log event to a processing function, and most events are not relevant. Which solution meets these requirements?

1. A Add a filter pattern to the subscription so only matching events are delivered to the processing function.
2. B Reduce the log group's retention period so fewer events accumulate.
3. C Add a second subscription filter so the volume is shared between two functions.
4. D Increase the processing function's memory allocation so it handles the volume more quickly.

175. An Amazon CloudFront distribution serves responses that differ depending on a request header the application sets, and users are receiving responses intended for other clients. Which solution meets these requirements?

1. A Add the header to the origin request policy without adding it to the cache key.
2. B Add that header to the cache key in the distribution's cache policy so responses are cached separately per header value.
3. C Reduce the distribution's default time to live so cached responses expire sooner.
4. D Disable caching on the distribution so every request reaches the origin.

176. A request traverses API Gateway, Lambda, and DynamoDB, and the team must identify which segment is slow. Which solution meets these requirements?

1. A Review the DynamoDB table's consumed capacity metrics.
2. B Review the API's access logs for total latency.
3. C Enable AWS X-Ray active tracing and instrument the SDK calls so each segment appears in the trace.
4. D Review the Lambda function's CloudWatch Logs for timing statements.

177. A Lambda function is failing intermittently, and the team must find the error message for the failed invocations. Which source should be examined?

1. A The function's configuration in the console.
2. B The function's reserved concurrency setting.
3. C The function's CloudWatch Logs log stream for the failed invocations.
4. D The CloudTrail record of the function's creation.

178. An application must emit a custom metric without a synchronous API call on every event. Which solution meets these requirements?

1. A Call PutMetricData for each event.
2. B Increase the log group's retention period.
3. C Write the metric as a structured log entry in the CloudWatch embedded metric format so CloudWatch extracts it asynchronously.
4. D Create a CloudWatch alarm on an existing metric.

179. A developer must search application logs across several Lambda functions for a correlation identifier. Which solution meets these requirements?

1. A Create a metric filter for the identifier.
2. B Download each log group and search locally.
3. C Run a CloudWatch Logs Insights query across the relevant log groups filtering on the identifier.
4. D Increase the log groups' retention periods.

180. An application's logs must be machine-parseable so fields can be queried individually. Which approach is appropriate?

1. A Emit logs as structured JSON with named fields.
2. B Emit logs as formatted prose with a consistent prefix.
3. C Emit logs with a fixed column width.
4. D Emit logs only on error.

181. A developer must add custom segments to an X-Ray trace to measure a specific block of application code. Which approach is appropriate?

1. A Add a log statement before and after the block.
2. B Enable active tracing on the function.
3. C Create a subsegment around the code block with the X-Ray SDK.
4. D Increase the X-Ray sampling rate.

182. An application must report a business event count that can be alarmed on. Which approach is appropriate?

1. A Write it to the application log only.
2. B Publish it as a CloudWatch custom metric with appropriate dimensions.
3. C Store it in a DynamoDB table.
4. D Include it in the API response.

183. An API's responses for a frequently requested resource rarely change. Which solution reduces backend load?

1. A Increase the API's throttling limit.
2. B Enable API Gateway caching for that resource with an appropriate time to live.
3. C Enable request validation on the API.
4. D Increase the backend's instance size.

184. A Lambda function is slower than expected, and profiling shows most time in CPU-bound work. Which adjustment is appropriate?

1. A Reduce the function's memory allocation to save cost.
2. B Increase the function's memory allocation, since CPU scales proportionally with memory.
3. C Increase the function's timeout.
4. D Increase the function's reserved concurrency.

185. An application makes many small writes to DynamoDB that could be grouped. Which approach improves efficiency?

1. A Use TransactWriteItems for all writes.
2. B Use BatchWriteItem to group the writes into fewer requests.
3. C Increase the table's provisioned write capacity.
4. D Use PutItem in a loop with a shorter timeout.

186. A Lambda function is throttled because downstream requests exceed a third-party API's rate limit. Which approach is appropriate?

1. A Increase the function's timeout so throttled calls complete.
2. B Set reserved concurrency on the function to cap its parallel executions below the downstream limit.
3. C Increase the function's memory allocation.
4. D Retry the downstream calls immediately.

187. A developer must identify which downstream call in a request caused an error. Which approach is appropriate?

1. A Examine the function's total duration.
2. B Examine the trace's subsegments for the request, which show each downstream call and its outcome.
3. C Examine the number of invocations.
4. D Examine the aggregate error rate metric.

188. An intermittent failure must be correlated across several services handling the same request. Which approach is appropriate?

1. A Propagate a correlation identifier through every service call and include it in each service's log entries.
2. B Compare the services' log timestamps.
3. C Increase every service's log verbosity.
4. D Review each service's logs independently.

189. A developer must determine whether a failure originated in the application or in an AWS service call. Which approach is appropriate?

1. A Examine the function's memory configuration.
2. B Examine the account's service quotas.
3. C Examine the application's total error count.
4. D Examine the trace for the request, since segment and subsegment errors distinguish application faults from downstream service faults.

190. An error occurs only under concurrent load and cannot be reproduced with a single request. Which approach is appropriate?

1. A Retry the single request several times.
2. B Run a load test that reproduces the concurrency and capture traces during it.
3. C Review the code for obvious defects.
4. D Increase the function's timeout.

191. An application must emit a metric with dimensions allowing it to be filtered by customer. Which approach is appropriate?

1. A Publish the metric with a customer dimension, keeping the dimension cardinality within practical limits.
2. B Publish a separate metric name per customer.
3. C Include the customer in the metric's log message only.
4. D Publish the metric without dimensions and filter downstream.

192. An application's traces must include a business identifier so specific transactions can be located. Which approach is appropriate?

1. A Include the identifier in the application's log only.
2. B Add the identifier as an annotation on the trace so it can be filtered on.
3. C Include the identifier in the response payload.
4. D Add the identifier as metadata on the trace.

193. An application must record structured context with every log entry without repeating it in each statement. Which approach is appropriate?

1. A Emit a separate log entry containing the context once per request.
2. B Store the context in a database keyed by request.
3. C Configure the logger with a context that is attached to every entry automatically.
4. D Concatenate the context into each log message.

194. An application must expose a health signal distinguishing a running process from a functioning service. Which approach is appropriate?

1. A Report healthy whenever the port is open.
2. B Implement a health endpoint that checks the application's critical dependencies before reporting healthy.
3. C Report healthy on a fixed schedule.
4. D Report healthy whenever the process is running.

195. A developer must measure how long a specific block of application code takes within a traced request. Which approach is appropriate?

1. A Measure the request's total duration.
2. B Increase the trace sampling rate.
3. C Create a custom subsegment around the block so its duration appears in the trace.
4. D Log the time before and after the block.

196. An application must record the outcome of each business transaction in a form that supports alerting. Which approach is appropriate?

1. A Write the outcome to the application log only.
2. B Include the outcome in the API response.
3. C Store the outcome in a database table.
4. D Emit a metric with a success or failure dimension so an alarm can evaluate the failure rate.

197. A DynamoDB query retrieves items and then discards most of them in the application. Which improvement is appropriate?

1. A Increase the table's provisioned read capacity.
2. B Add a filter or refine the key condition so fewer items are read, since filters apply after the read is charged.
3. C Cache the results in the application.
4. D Increase the application's memory allocation.

198. An application makes the same read request to a downstream service many times within a short period. Which improvement is appropriate?

1. A Increase the application's concurrency.
2. B Cache the response for a period appropriate to how quickly the data changes.
3. C Increase the downstream service's capacity.
4. D Retry the requests with backoff.

199. An application must record the duration of each external call for later analysis. Which approach is appropriate?

1. A Record the total request duration only.
2. B Record the number of external calls made.
3. C Emit the duration as a metric with a dimension identifying the external service.
4. D Log the start and end timestamps for each call.