Oracle ' s microservices design guidance explicitly identifies responsive behavior, backward compatibility, flexible communication, and idempotency as important requirements. Flexible communication doesnotrequire every microservice to use one synchronous protocol. Oracle states that each microservice can specify the protocol used between clients and the API gateway as well as the protocol used between individual microservices. This flexibility supports loose coupling and allows synchronous or asynchronous communication where appropriate. Oracle also states that microservices should return a response even when the service fails, API changes should remain backward-compatible with existing clients, and repeated calls should produce the same outcome when idempotency is required. Consequently, option C directly contradicts Oracle ' s documented flexible-communication principle and is the NOT valid statement.

Explanation

The correct statements about deleting a Kubernetes cluster are: If you change the auto-generated name of a worker node and then delete the cluster, the renamed worker node is not deleted. Changing the name of a worker node does not affect its deletion when the cluster is deleted. The cluster deletion process does not consider the renamed worker nodes and will delete all worker nodes associated with the cluster. Upon deleting a cluster, no other resources created during the cluster creation process or associated with the cluster (such as VCNs, Internet Gateways, NAT Gateways, Route Tables, Security Lists, Load Balancers, and Block Volumes) are deleted automatically. These additional resources are not automatically deleted when the cluster is deleted. You need to manage the deletion of these resources separately, if desired. Therefore, the correct statements are that the renamed worker nodes are not deleted when the cluster is deleted, and other associated resources are not automatically deleted when the cluster is deleted.

The LATEST cursor type positions the consumer at the most recent point in the stream so that it receives only messages published after the cursor is created. Oracle explicitly states that a cursor based on the most recent message allows consumption only of messages subsequently added to the stream. TRIM_HORIZON has the opposite behavior: it begins with the oldest message still available inside the stream ' s retention period. AFTER_OFFSET begins immediately after a supplied partition offset, while AT_TIME begins from a specified timestamp. Because the requirement is to ignore all historical telemetry already retained in the partition and consume only newly arriving records, LATEST is the correct cursor type for the CreateCursor operation.

Explanation

The strategy that involves exposing new functionality or features to only a small set of users is called Canary Deployment. Canary deployment is a technique used in software development and deployment where a new version of an application or feature is released to a small subset of users or a specific group of servers. This allows for testing and gathering feedback on the new functionality in a controlled and limited environment before making it available to a wider audience. In a canary deployment, a small portion of the traffic is routed to the new version while the majority of the traffic still goes to the stable version. This allows for monitoring and evaluation of the new functionality in real-world conditions while minimizing the impact of any potential issues or bugs. If the new version performs well and meets the desired criteria, it can then be gradually rolled out to a larger user base or all servers. By exposing the new functionality or features to a small set of users initially, canary deployment helps in identifying any unforeseen issues, gathering feedback, and ensuring the stability and reliability of the application before a full deployment.

Explanation

The option that is NOT available when viewing the Application metrics in the Oracle Cloud Infrastructure (OCI) Console is: " The number of retries made by the function before failing due to an error. " When viewing the Application metrics in the OCI Console for Oracle Functions, you can typically see metrics related to the performance and usage of your functions. These metrics provide insights into how your functions are performing and being utilized. The following metrics are usually available: The number of requests to invoke a function that failed due to throttling: This metric indicates the number of requests that were not processed by the function due to reaching the configured concurrency limit or throttling settings. The length of time a function runs for: This metric represents the duration of each function invocation, measuring the time it takes for the function to complete its execution. The number of requests to invoke a function that failed with an error response: This metric counts the number of requests that encountered an error during the function invocation, resulting in a failed response. However, the number of retries made by the function before failing due to an error is not typically available as part of the Application metrics in the OCI Console. The retries made by the function are usually handled at the invoker level, and the specific details of retries may not be captured as part of the application-level metrics. It ' s important to note that the availability of metrics and their specific details may vary depending on the version and configuration of Oracle Functions and the monitoring setup. It is recommended to refer to the Oracle Functions documentation and consult the official documentation for accurate and up-to-date information on available metrics.

Oracle Monitoring documents groupBy(x) specifically for aggregating metric streams according to a shared dimension. Oracle ' s own fault-domain example uses the exact expression CpuUtilization[1m].groupBy(faultDomain).mean(). The resulting chart contains one aggregated metric stream for each fault domain, which means a compartment containing three fault domains produces three lines rather than one line per Compute instance. grouping() combines all selected metric streams into a single aggregate instead of retaining separate fault-domain groups. filter(faultDomain) is not the appropriate MQL grouping operation, and split() is not the documented operator for this requirement. Because the scenario requires mean CPU utilization separately aggregated for each value of the faultDomain dimension, option B exactly matches Oracle ' s documented MQL pattern.

Explanation

The correct answers are: Set up the following dynamic group for your function ' s OCID: Name: read-file-dg Rule: resource.id = " ocid1.fnfunc.oc1.phx.aaaaaaaakeaobctakezjz5i4ujj7g25q7sx5mvr55pms6f4da " Set up a policy with the following statement to grant read access to the bucket: Statement: allow dynamic-group read-file-dg to read objects in compartment qa-compartment where target.bucket.name = ' input-bucket ' Explanation: To implement the use case of reading a JSON file object from an Object Storage bucket using Resource Principals with Oracle Functions, you need to configure the following: Create a dynamic group named " read-file-dg " and associate it with your function ' s OCID. This dynamic group helps identify the function as a member of the group for policy enforcement. Create a policy that grants read access to the bucket. The policy statement should allow the dynamic group " read-file-dg " to read objects in the compartment " qa-compartment " and specify the target bucket name as " input-bucket " . This policy ensures that the function has the necessary permissions to access the specified bucket. By setting up the dynamic group and policy, you ensure that the function, as a member of the dynamic group, has the required read access to the specified Object Storage bucket in the specified compartment.

Oracle identifies distributed transaction consistency and observability as major concerns when moving from monoliths to microservices. The Saga pattern addresses cross-service business transactions by dividing them into local transactions and using events plus compensating actions, avoiding long-duration distributed locks. Oracle also emphasizes unified observability for microservices, including metrics, logs, and traces, because failures become harder to diagnose when calls cross multiple networked services. OCI does not automatically create a global application transaction across compartments, Object Storage is not a transactional relational database replacement, and microservices are not restricted to synchronous REST communication. Oracle explicitly supports asynchronous messaging and event-driven designs for loose coupling. Therefore, A and B correctly describe realistic migration challenges and strategies.

The supplied Word answer key requires correction here. Current Oracle Queue documentation explicitly states that adead letter queue is automatically created when a queue is created. Therefore, option C is correct and option B is not. Oracle also provides the Maximum delivery attempts setting, which controls how many unsuccessful deliveries can occur before a problematic message is transferred to the DLQ. Setting that value to 5 implements the requested retry threshold. No separately provisioned target queue is required, consumer groups are optional and unrelated to enabling DLQ functionality, and administrators do not create an IAM policy authorizing a source queue resource to write to its automatically associated DLQ. Consequently, the verified current answer isA, C, not the Word file ' s A, B.

Oracle explicitly states that theNONEparser type is required when configuring the Unified Monitoring Agent even when no particular parsing of the incoming log structure is desired. The agent configuration still requires a parser declaration so it knows how the input should be handled. NONE allows the log content to be collected without applying structural parsing such as JSON, SYSLOG, regular-expression, or Apache-style parsing. Oracle ' s SDK model also lists NONE as a supported parserType value. RAW, DEFAULT, and UNSTRUCTURED are not the documented parser-type identifiers for this purpose. Consequently, administrators configuring custom logs that should be ingested without parsing must explicitly select or specify NONE, making C the correct answer.