Free C9050-041 exam questions in PDF & AI Tutor

Menna Nawar on: May 01, 2026
Thanks to God, I passed my exam with the help of this amazing website in April 16.
Ash-Shaykh Zayid, Egypt

Community-Verified AI Explanation on: May 01, 2026
Question 16:
Question 16: During the ML process, when should you review evaluation metrics?

Correct answer: D) After you test a model on the validation data.

Why:

You need a trained model to generate predictions on the validation data in order to compute evaluation metrics (e.g., accuracy, precision, recall).

Reviewing metrics before training, or after only cleaning data or choosing the model, doesn’t produce meaningful evaluation results.

In practice, you train on training data, evaluate on validation data, and then review the metrics to decide if you need to tune hyperparameters or compare different models.

Mokopane, South Africa

Community-Verified AI Explanation on: May 01, 2026
Question 6:

Correct answer: D

Why: The Power Query function to get per-column statistics (Count, Average, Null count, Distinct count, Standard deviation, etc.) is Table.Profile. It returns a profile table with those metrics for each column.

Syntax: Table.Profile(table as table, optional additionalAggregates as nullable list) as table

What it provides: minimum/maximum, average, standard deviation, count, null count, and distinct count for each column.

Why the other options aren’t correct:

- Table.Schema: returns column metadata (name, type), not statistics. - Table.View: returns a view of the table, not statistical summaries. - Table.FuzzyGroup: used for fuzzy grouping of rows, not column statistics.
Mumbai, India

Community-Verified AI Explanation on: May 01, 2026
Question 5:
Here’s why the three you should pick are B, C, and E.

To get min and max per column, you first need column profiling enabled. This collects statistics for each column, so you select Enable column profile (B).

Once profiling is on, you need to view those statistics. The min and max are shown in

Mumbai, India

Community-Verified AI Explanation on: April 30, 2026
what is api and json

API (Application Programming Interface): A set of rules and endpoints that let one software application talk to another. It defines how to request services, pass data, and receive responses. In web apps, APIs are often RESTful and use HTTP requests.

JSON (JavaScript Object Notation): A lightweight, text-based data format used to represent structured data. It uses key-value pairs and arrays, e.g. { "name": "Alex", "role": "admin" }. It’s human-readable and easy for machines to parse.

How they relate:

Many APIs exchange data in JSON format over HTTP/HTTPS. A client sends a request to an API endpoint and gets back JSON data.

Simple example:

API call: GET https://api.example.com/users/123

Response (JSON): { "id": 123, "name": "Alice", "email": "alice@example.com" }

Security notes:

Use encryption with TLS/HTTPS.

Authenticate (API keys, OAuth 2.0) and authorize requests.

Validate inputs and sanitize JSON to prevent injection.

If you want, I can walk through more examples or relate these to specific Security+-style questions.
East Northport, United States

Community-Verified AI Explanation on: April 30, 2026
Question 17:
Short answer: The correct option is E — only new data added to the eventhouse.
Why:

Enabling OneLake availability for an eventhouse typically enables incremental replication to OneLake. It does not automatically copy all existing data at activation.

After you enable it, OneLake will start copying data that gets added

New York, United States

Community-Verified AI Explanation on: April 30, 2026
Question 16:

The correct answer is: B) a deployment pipeline

Why:

- A deployment pipeline in Fabric is designed to move assets (like warehouses) between workspaces and to validate them as part of the process. - It provides a built-in validation step to detect invalid references (dependencies, missing assets, broken links) before the deployment, helping you catch issues with minimal extra development. - Using a deployment pipeline minimizes manual effort compared with building custom checks in a database project, a Python script, or a T-SQL script.

How it works (high level):

- Create a pipeline with Source: Workspace1 and Target: Workspace2. - Run the validation/deploy flow to surface any invalid references. - Fix issues and redeploy as needed.

Contrast with other options:

- A database project, Python script, or T-SQL script would require you to implement and maintain custom validation logic, increasing development and maintenance work.
New York, United States

Community-Verified AI Explanation on: April 30, 2026
Question 8:
The correct answer is D: echo "profile-${num}-$name"
Why:

In shell scripting, variable expansion uses $var. Using braces, as in ${num}, clearly marks the end of the variable name. This prevents any ambiguity if the text immediately after could be part of a variable name.

The other options fail for these reasons:

- A: $(num) would try to execute a command named num and substitute its output. - B: Single quotes prevent expansion entirely. - C: Backticks around Snum attempt command substitution and are syntactically incorrect here (and Snum isn’t a valid command).

While echo "profile-$num-$name" often works, using ${num} is safer when the surrounding text could be adjacent to a character that starts/ends a variable name (e.g., hyphens, letters, etc.).

Innisfil, Canada

Community-Verified AI Explanation on: April 30, 2026
Question 1:
Correct answer: a, b, and c (the 4th option).
Why:

a) True. In Form Designer, dragging a field from the Fields tab to the form adds it to the form layout without changing the underlying table.

b) True. Dragging a data type from the Field Types tab to the form creates a new field on the form’s table (adds a new column) and places it on the form.

c) True. Hovering a field enables action buttons; clicking Delete removes the field from the form view only (the table definition remains).

d) False. Sections aren’t dragged from the Field Types tab. To add a section, use the Add Section option on the toolbar or the right-click menu.

Key takeaway: you can modify the form layout independently of table schema (a, b, c true), and sections are added via a separate Add Section action, not by dragging from Field Types.
New York, United States

Community-Verified AI Explanation on: April 30, 2026
Question 62:
Question 62 asks what the CSIRT should do next after isolating a possibly compromised VM.

The correct answer is: A — Take a snapshot of the compromised server and verify its integrity.

Why:

After containment, the priority is evidence preservation. A VM snapshot captures the exact state (memory and disk) at that moment, enabling forensic analysis later.

Verifying integrity (e.g., hashes, chain of custody) helps ensure the evidence hasn’t been tampered with and can be used in investigations.

Why not the others:

B: Restoring the server would remove malware but also destroy or alter forensic evidence.

C: Involving a government agency isn’t automatically required; it’s typically a policy decision or legal requirement.

D: Attribution is part of investigation but not the immediate next step; initial focus is evidence collection and containment.

Tip: In virtual environments, prioritize capturing volatile data (RAM) via a snapshot or memory capture, and avoid rebooting or making changes that could contaminate the evidence.
London, United Kingdom

Community-Verified AI Explanation on: April 30, 2026
Question 93:
I’m not able to see the image, but based on the provided reference, Question 93 is about creating features in the data science process.

Correct answer concept: Feature engineering (creating features from raw data).

Why it’s correct: In ML, feature engineering is the process of using domain knowledge to derive new features from existing data that help the model learn patterns better. This often includes:

- Transformations (normalization, log, scaling) - Encoding categorical variables (one-hot, ordinal) - Interaction or aggregate features (ratios, differences, counts) - Time-based features (year, month, day of week, hour) - Handling missing values and creating robust features

Why this matters for AI-900: It aligns with the data preparation and feature creation concepts covered in Azure ML and the data science workflow.

If you can share the exact sentence you’re choosing from, I can map it to the appropriate term and explain why A fits (or why another option might be correct).
Noida, India

Community-Verified AI Explanation on: April 30, 2026
Question 1:
Question 1 asks which role is needed for a user to enable Azure AD Privileged Identity Management (PIM) in a new subscription.

Correct answer: Global administrator (A)

Why:

- Enabling PIM is a directory-wide configuration task. You must sign in to the Azure portal as a Global Administrator with an organizational account to enable PIM for the directory. - Lower-privilege roles (Security administrator, Password administrator, Compliance administrator) do not grant the permissions required to enable PIM. - After PIM is enabled, you can assign other users to manage privileged roles (for example, Privileged Role Administrator), but enabling the feature itself requires Global Administrator access.

Quick note on the other options:

- Security administrator, Password administrator, Compliance administrator: do not provide the necessary permission to enable PIM.
Derby, United Kingdom

Community-Verified AI Explanation on: April 30, 2026
Question 2:
The correct approach is: Randomly split the data into rows for training and rows for evaluation.
Explanation:

Use the Split Data module with the Split Rows option to create separate training and evaluation (test) sets.

You can specify the percentage for each split (e.g., 70/30, 80/20) and choose to randomize which rows go into each set.

Stratified sampling can be used to preserve the class distribution in each split.

Why the other options are wrong:

A: Training should use features and labels together in the training set; evaluation needs the ground-truth labels for comparison, not just features.

C: This reverses the roles of labels and features and is not a valid data split.

D: Splitting by columns would split features, not instances, which is not the standard train/eval split.

Lagos, Nigeria

Community-Verified AI Explanation on: April 30, 2026
Question 36:

Answer: False

Why:

- Snowflake separates storage from compute. Tables live in centralized storage and are not bound to the warehouse that loaded them. - Any virtual warehouse with the proper privileges (e.g., SELECT on the table) can query the table, even if it didn’t load the data. - This allows multiple warehouses to query the same table concurrently without being tied to the loading warehouse.
Aligarh, India

Community-Verified AI Explanation on: April 29, 2026
Question 25:
Correct answer: A
Why A is best:

The scenario is about accommodating a team member’s new personal commitments that affect the originally agreed working hours. In Agile, you can adapt cadence to match team capacity. Changing the iteration duration is a focused, low-risk way to realign schedule while still delivering value.

Why the other options are less suitable:

B (Validate goals with the customer): Important, but it doesn’t address the capacity/scheduling issue.

C (Invite the customer to daily meetings): Unnecessary for this internal constraint and could create scope/cadence issues.

D (Help redefine the MVP): Moves scope rather than addressing the team’s capacity and cadence problem.

Practical steps you could take:

Recalculate team capacity and adjust the next iteration duration or sprint length accordingly.

Replan the iteration backlog to fit the new cadence.

Communicate changes with the team and stakeholders; ensure alignment on expectations and value delivery.

Newnan, United States

Community-Verified AI Explanation on: April 29, 2026
What is true about Python packages? (Select two answers) the__name__variable content determines the way in which the module was run a package can be stored as a tree of sub-directories/sub-folders __pycache__is the name of a built-in variable hashbang is the name of a built-in Python function
Correct choices: A and B

A: The value of the built-in variable __name__ indicates how the module was run. If you execute a file as a script, __name__ is '__main__'; if it’s imported, __name__ is the module’s name.

B: A package can be stored as a tree of sub-directories/sub-folders. Packages are directories containing modules (and possibly subpackages).

Why the others are false:

__pycache__ is a directory for cached bytecode, not the name of a built-in variable.

A hashbang (#!/...) is a Unix script header, not a built-in Python function.

Freiburg Im Breisgau, Germany

Community-Verified AI Explanation on: April 29, 2026
Question 11:

Correct answer: E. Run the Set-AzureStaticVNetIP PowerShell cmdlet.

Why: You want to assign static internal IPs to two VMs in a single Azure VNet subnet. In the classic Azure (Azure Service Management) approach, you set a static private IP with Set-AzureStaticVNetIP and then apply the change with Update-AzureVM. Example:

Get-AzureVM -ServiceName StaticDemo -Name VM2 | Set-AzureStaticVNetIP -IPAddress 192.168.4.7 | Update-AzureVM

How it works:

- Set-AzureStaticVNetIP designates a private (internal) IP for a VM that was already created. - After updating, the VM may restart; the static IP is then assigned.

Why other options aren’t correct:

- New-AzureRMVMConfig creates a new VM config (not for setting an existing VM’s IP). - Set-AzureSubnet adjusts subnet settings, not per-VM IPs. - Modifying VM properties in the portal or inside Windows (Network and Sharing Center) does not configure Azure-level static private IPs for the VM’s NIC.
Innisfil, Canada

Keketso moselantja on: April 29, 2026
This has improved my level of thinking I now act like a Leader
Maseru, Lesotho

Community-Verified AI Explanation on: April 29, 2026
Why Withdraw will not work?

Withdraw will not work because Withdrawn permanently retires the circumstanced instance from rule resolution. It is not easily re-enabled and effectively removes the circumstance from being used.

In this scenario, you want to disable the Silver circumstance in the new version while keeping the rule in the system for auditing and potential re-use. Setting the circumstanced instance to Blocked achieves this: it’s not used in resolution, but remains available to re-enable later or reference.

If you used Withdrawn, you would be removing the circumstance entirely, which is more permanent and would require recreating or re-adding the circumstanced rule later if needed.

Mumbai, India

Community-Verified AI Explanation on: April 29, 2026
Question 145:
There are no more questions on this page beyond Q145.
If you’d like, I can generate new practice questions (e.g., Q146–Q150) with concise answers and explanations on topics like:

provider version constraints and lock file

remote state and backends

Terraform Cloud features (workspaces, VCS integration)

writing a custom provider or Plugin Framework basics

Sentinel policy structure

Would you like me to create them now? If you have a preferred topic, tell me and I’ll tailor the questions.
Hyderabad, India

Community-Verified AI Explanation on: April 29, 2026
Question 1:

Correct answer: Transfer (option B)

Why: Purchasing cyber insurance is a risk response that shifts the financial impact of a risk from the organization to the insurer. This is risk transfer.

Why not the others:

- Mitigate: reduce likelihood or impact of the risk, not transfer it. - Avoid: stop the activity that creates the risk entirely. - Accept: acknowledge the risk and proceed without any action or cost.

Quick example: If a potential data breach could cost $X, insurance helps cover those costs, rather than the company paying the full amount itself.

Lancaster, United States

BISSS on: April 29, 2026
Will see after exam:)
Dudley, United Kingdom

Community-Verified AI Explanation on: April 29, 2026
Question 523:
Correct answers: B, C, D
Explanation:

B: Telnet access should be authenticated via AAA. This centralizes authentication and allows ACLs/policies to govern access, not just device-local credentials.

C: To give different levels of authority, configure different privilege levels for each administrator’s account. This controls what commands they can run.

D: With AAA, you typically configure multiple user accounts (three in this case) and their passwords in the AAA database (local AAA or external AAA server).

Why not A:

A states each administrator uses a different public IP to run Telnet. IP address alone does not grant authority or control access levels. You can restrict by IP, but that does not provide per-user authentication and authorization. The correct approach is to use AAA with distinct user accounts and privilege levels (B, C, D).

Johannesburg, South Africa

Community-Verified AI Explanation on: April 29, 2026
Question 8:

Correct answer: A. Public cloud

Why: The public cloud provides on-demand, scalable computing resources across vast infrastructure without you purchasing or provisioning new hardware. You can scale up or down quickly as needed and pay only for what you use.

Why the other options aren’t correct:

- Containers: packaging/runtime mechanism, not a resource provider. - Private cloud: runs on organization’s own hardware or hosted privately; capacity is limited by that hardware. - Microservices: architectural pattern for building applications, not a resource provisioning model.

Key concept: Public cloud enables elasticity and rapid growth without capital expenditure, aligning with the exam’s emphasis on scalable, flexible cloud resources.

Mumbai, India

Community-Verified AI Explanation on: April 29, 2026
Question 100:
Question 100 asks which method of the DataDriftDetector class to use to compare dataset1 (baseline) with dataset2 (new data).

Correct answer: backfill

Why:

- backfill runs are used to see how data changes over time. It allows you to analyze drift across historical or sequential data, which fits comparing baseline vs. new data over time. - Other options: - run starts a drift detection run for the current data pair. - get retrieves the results of a drift run. - update updates the detector with new data for future checks.
Key concept: use a backfill run when you want to observe how drift evolves over time between baseline and new data, rather than just performing a single snapshot check.
Porto, Portugal

Gabriel on: April 29, 2026
studying for dp-100
Porto, Portugal

sanmar on: April 29, 2026
questions are vesy useful
Muscat, Oman

Community-Verified AI Explanation on: April 29, 2026
Question 60:

Correct answer: Implement moderation APIs.

Why: Moderation APIs screen images for policy-violating content (inappropriate, explicit, violence, etc.) before they are returned by the chatbot, directly preventing unwanted images.

Why the others are not sufficient:

- Retrain the model with a general public dataset — does not guarantee removal of inappropriate images and may not cover all policy gaps. - Perform model validation — assesses overall performance, not specifically filtering or blocking inappropriate outputs. - Automate user feedback integration — helps improve over time but does not prevent bad images from being shown in the first place.

How to use generally: integrate the moderation step in the image output pipeline; block or replace content flagged above threshold; log decisions and optionally allow user feedback.

Cape Town, South Africa

Community-Verified AI Explanation on: April 29, 2026
Question 55:

Correct answer: A — Generative adversarial network (GAN)

Why:

- GANs are designed to generate new samples that resemble the training data by training a generator and a discriminator in an adversarial process. This makes them a primary choice for synthetic data generation.

Why the other options aren’t suitable here:

- XGBoost is for predictive modeling on structured data, not for generating new data. - Residual neural network is a network architecture, not specifically a data-generation model. - WaveNet specializes in generating audio waveforms, not generic synthetic data across domains.

Quick note: GANs are widely used for creating synthetic data across various modalities (images, tabular, text), but consider privacy and bias considerations when deploying synthetic data.

Cape Town, South Africa

Community-Verified AI Explanation on: April 29, 2026
Question 50:

Correct answer: A — Purchase Provisioned Throughput for the custom model.

Why:

- In Amazon Bedrock, using a custom (private) foundation model requires provisioning throughput to allocate the necessary compute capacity for serving requests. This step enables Bedrock to route and process inferences for your custom model.

Why the other options aren’t required here:

- Deploying the custom model in a SageMaker endpoint isn’t needed for Bedrock usage, as Bedrock handles inference through its own pathway. - Registering the model with the SageMaker Model Registry isn’t a prerequisite for Bedrock integration. - Granting access to the custom model in Bedrock isn’t the mandatory step described for enabling usage; provisioning throughput is.
Cape Town, South Africa

Naveen on: April 29, 2026
Aws sample question
Hyderabad, India

Community-Verified AI Explanation on: April 28, 2026
microsoft defender portal is not the same as purview compliance portal
You’re right: they are not the same portal. The Microsoft Defender portal and the Purview (compliance) portal are separate experiences.
For this task, the audit data can be accessed from either location:

Purview compliance portal (Audit):

- Navigate to Audit, set the date range, Activities = “Shared file, folder, or site,” Workload = OneDrive, Run, then export.

Microsoft Defender portal (Audit):

- Open the Audit/Activity log area, filter by similar criteria (Sharing activities, OneDrive), run the query, and export.
Key point: both portals expose the same underlying audit data, but the navigation paths differ. If you want, I can outline exact, up-to-date steps for each portal.
Montreal, Canada

Piotr on: April 28, 2026
Please let me know where is more questions? I received only 300 but you mensioned there are more than 800
Gdansk, Poland

Community-Verified AI Explanation on: April 28, 2026
Question 36:

Correct answer: IDS

Why:

- An IDS (Intrusion Detection System) monitors and analyzes traffic for suspicious activity and generates alerts, usually in a passive, out-of-band position (e.g., on a SPAN port or TAP). This means it can analyze attacks from the internet without being in the data path, so it doesn’t add latency. - An IPS (Intrusion Prevention System) is inline and can block or modify traffic. While it provides active protection, it can introduce processing delays and affect latency. - A Firewall filters traffic based on policies, which can add some latency but is not primarily an analysis tool. - A Load balancer distributes traffic and isn’t used for attack analysis.
So, for observing attacks without impacting latency, an IDS is the appropriate choice.
Tadcaster, United Kingdom

Lawrence James on: April 28, 2026
I find it very informative. the questions are laid out easy which means you can do several questions extremely fast.
Tadcaster, United Kingdom

Community-Verified AI Explanation on: April 28, 2026
Question 55:

Correct answer: A. Code repositories.

Why this is correct:

- Developers sometimes commit credentials (API keys, access tokens) into code or config files in code repositories. If those keys are exposed (even in private repos or via history), cloud services can be accessed by anyone who gets the keys.

Why the other options are less likely as the common source:

- Dark web: a place for exposed credentials after leakage, not the primary cause of inadvertent leakage in cloud setups. - Threat feeds: provide indicators, not the root cause of credential leakage. - State actors: target breaches, not the typical inadvertent leakage source. - Vulnerability databases: catalog vulnerabilities, not credentials.

Key concepts and best practices:

- Use secret management (e.g., AWS Secrets Manager, Azure Key Vault) instead of hard-coded secrets. - Implement secret scanning in IDEs/CI (e.g., detect secrets in commits). - Rotate credentials, enforce least privilege, and use short-lived tokens or roles.
Ikeja, Nigeria

Community-Verified AI Explanation on: April 28, 2026
Question 52:

Correct answer: D. Block the URL shortener domain in the web proxy.

Why this is correct:

- The emails use a single URL shortener to hide the final (dead) destination. Blocking that shortener domain at the web proxy prevents users from reaching any of those shortened links, stopping the threat from propagating across multiple inboxes.

Why the other options are less effective:

- A: Blocking all subject lines is impractical; attackers can change subjects easily. - B: Sending the dead domain to a DNS sinkhole aids analysis but does not prevent users from receiving or clicking links in the moment. - C: Quarantining all emails is thorough but can disrupt legitimate mail and is slower than a targeted block.

Additional tips:

- After blocking, monitor for new shorteners being used and update filters. - Consider user awareness training and additional URL reputation controls.
Ikeja, Nigeria

Shailendu Dwivedi on: April 28, 2026
It is really helpful
Noida, India

Community-Verified AI Explanation on: April 28, 2026
Question 102:

Correct answer: A. Further review closed unactioned alerts to identify mishandling of threats.

Why this is best:

It validates management’s claim by gathering evidence. Closed alerts without action may indicate weak threat response, misclassification, or policy gaps.

A risk-based, evidence-led approach helps determine the actual impact and root causes before taking broader actions.

It informs whether further actions are needed (policy changes, training, process improvements) rather than jumping to conclusions or omitting the finding.

Why not other options:

Reopening unactioned alerts and reporting to the audit committee moves too quickly to formal escalation without first understanding scope and impact.

Recommending policy/training improvements is forward-looking but not the first step without evidence from a detailed review.

Omitting the finding is inappropriate when there is potential risk demonstrated by unaddressed alerts.

Riyadh, Saudi Arabia

Community-Verified AI Explanation on: April 28, 2026
Question 28:

Correct answer: A)

Why: Azure Government is a separate, sovereign cloud region intended for U.S. government needs. It serves:

- U.S. government agencies (federal, state, local) - approved government contractors/partners who support government workloads
It provides compliance and data residency specific to government workloads (e.g., FedRAMP High, DoD alignment). It is not available to general public or non-government customers.
Bengaluru, India

Community-Verified AI Explanation on: April 28, 2026
Question 25:

Correct answer: C) Configure a Point-to-Site (P2S) VPN

Why: A Point-to-Site (P2S) VPN lets individual remote client computers securely connect to an Azure VNet. This is exactly what remote users need to access VNet1 from outside Azure.

Why not the others:

- Site-to-Site (S2S) VPN) connects on-premises networks to Azure, not individual remote users. - VNet-to-VNet VPN connects two VNets, not client devices. - DirectAccess is not the typical solution for Azure VNet access from multiple remote clients. - Multi-Site VPN is for connecting multiple sites, not individual client machines.

Quick how-to (high level):

- Create a VPN gateway in Azure and enable P2S. - Configure authentication (certificate-based or RADIUS). - Download and distribute the VPN client configuration to users. - Users install the client and connect to the Azure VNet to access VMs.
Bengaluru, India

Community-Verified AI Explanation on: April 28, 2026
Question 24:

Correct answer: Yes.

Why:

The goal is to reduce administrative effort when deploying a large number of VMs weekly (mixed OS: Windows Server 2016 and Ubuntu Linux).

Azure DevTest Labs is designed to simplify and accelerate the creation, management, and teardown of labs with standardized VM images and templates.

Key benefits that meet the goal:

- Self-service, reusable VM templates for both Windows and Linux. - Quotas, policies, and cost controls to prevent over-provisioning. - Artifacts to automate software installation/configuration. - Auto-shutdown/idle timeout to reduce costs and admin overhead.

These features collectively reduce manual provisioning, configuration, and deprovisioning tasks, aligning with the requirement to minimize administrative effort.

Bengaluru, India

Community-Verified AI Explanation on: April 28, 2026
Question 13:

Question 13 asks if using Azure Cosmos DB meets the goal of building, testing, and deploying predictive analytics for an AI solution.

The correct answer is No (B).

Why:

Cosmos DB is a database service, not a platform for developing or deploying ML models.

To build/test/deploy predictive analytics, you should use Azure Machine Learning (or a similar ML service) to create models, run experiments, and deploy endpoints. Data can be stored in Cosmos DB, but the ML workflow is handled by Azure Machine Learning.

Therefore, the proposed solution does not meet the goal.

Bengaluru, India

Community-Verified AI Explanation on: April 28, 2026
Question 2:

Question 2 asks you to pick a storage redundancy option that stores data in multiple locations and allows reading from the secondary site.

Correct answer: B. Read-only geo-redundant storage (RA-GRS).

Why:

- RA-GRS replicates data to a secondary region (data in separate geographic locations). - It provides read access to the data in the secondary location (read-only), meeting the requirement to read from both primary and secondary. - It is an opt-in feature that requires geo-replication.

Quick contrast:

- LRS (Locally Redundant Storage) keeps data in one data center only. - ZRS (Zone-Redundant Storage) replicates across zones within a single region. - GRS (Geo-Redundant Storage) replicates to a secondary region but does not inherently provide read access unless configured as RA-GRS.

Takeaway: When you need cross-region replication with read access to the secondary, choose RA-GRS.

Bengaluru, India

Community-Verified AI Explanation on: April 28, 2026
Question 118:

The correct action to raise service level from 95% to 97% is Increasing safety stock (Option A), not Decreasing safety stock.

Why:

Safety stock buffers against demand and supply variability. Increasing it reduces stockouts, which directly raises the service level.

Expanding replenishment lead time can worsen service level unless you also increase safety stock; it generally isn’t the best lever.

Holding more cycle stock increases total inventory but doesn’t reliably improve service level; it ties up capital without addressing variability.

Decreasing safety stock would typically lower the service level.

Note: The provided answer label (“Answer(s): C”) appears to be a misprint. The explanatory principle in the material confirms that higher safety stock increases customer service levels.
Plymouth, United Kingdom

Mathi on: April 28, 2026
Helped me understand AWS services
Chennai, India

Community-Verified AI Explanation on: April 28, 2026
Question 10:

Correct answer: D (4)

Explanation:

APIKit generates flows to implement each API operation defined in the RAML.

For every operation, APIKit creates a private flow that contains the internal logic to fulfill that operation.

There is typically one public flow per resource that routes requests to the appropriate private flow.

Therefore, if the RAML specifies four operations, APIKit will generate four private flows.

San Jose, United States

Gurdeep on: April 27, 2026
Ekdam Zabardast! All questions real and from the actual exam.
India

Community-Verified AI Explanation on: April 27, 2026
Question 5:
Correct answer: D. Authentication
Explanation:

The Remote Access panel in the User Activity dashboard relies on events related to user authentication (e.g., remote logins). Those events come from the Authentication data model.

If the panel isn’t showing the most recent hour, the underlying cause is often that scheduled searches feeding the Authentication data model were skipped, so new data wasn’t processed.

What to check:

In Splunk: go to Settings > Data models and open the Authentication data model. Ensure it’s active and its population/accelerations are running.

Check the scheduled searches that feed this data model for any Skipped status or recently failed runs (Saved Searches related to the Authentication model or ES user activity).

Look at last run times for these searches (e.g., via _internal or Saved Searches UI) to confirm they’re executing as expected.

If searches were skipped due to resource constraints, consider adjusting the schedule, reducing scope, enabling summary indexing, or increasing search head capacity.

Reason this matters: the panel’s data is pulled from Authentication data; skipped searches prevent the latest data from appearing.
Kollam, India

Community-Verified AI Explanation on: April 27, 2026
Question 63:
Answer: B — A packet capture tool was used to steal the password.

Why: Telnet transmits credentials in cleartext. If Telnet was regularly used to log in, an attacker on the same network could capture the username/password with a packet sniffer, and the malware could then use those credentials to access the server.

Why not the others:

- A spraying attack: would target many accounts, not specifically tied to the observed Telnet use. - A remote-access Trojan: could install malware, but the clue points to credential theft via sniffing, not a Trojan. - A dictionary attack: would brute-force credentials, not rely on Telnet’s plaintext transmission.
Mitigation: Disable Telnet; use encrypted protocols like SSH; enforce strong authentication and network monitoring.
Cartersville, United States

IBM C9050-041 Exam Questions
Programming with IBM Enterprise PL/I (C6030-041) (Page 26 )

C9050-041 Exam Discussions & Posts (Share your experience with others)

IBM C9050-041 Exam Questions Programming with IBM Enterprise PL/I (C6030-041) (Page 26 )

C9050-041 Exam Discussions & Posts (Share your experience with others)

IBM C9050-041 Exam Questions
Programming with IBM Enterprise PL/I (C6030-041) (Page 26 )