🔐 Enterprise Administration

Publishing a report is only half the battle. Once it's in the cloud, you must ensure the data stays up-to-date and, more importantly, that sensitive data doesn't fall into the wrong hands.

⚡ Performance Optimization

Before publishing, ensure your report is optimized:

• Performance Analyzer: A tool in Desktop (under the View ribbon) that records exactly how many milliseconds every visual takes to load. Use this to identify slow visuals.
• Query Folding: When Power Query translates your cleaning steps directly into a native SQL query, pushing the workload back to the database. This drastically improves refresh performance!
• DirectQuery Optimization: If using DirectQuery, use Star Schemas and avoid bidirectional filters to ensure the database isn't overloaded.

🛡️ Data Security: Row-Level Security (RLS) and OLS

Imagine you build a global Sales Report. You don't want the Manager of Germany to see the sales numbers for France. Do you build two separate reports? No! You use Row-Level Security (RLS).

There are two primary ways to secure data:

1. Row-Level Security (RLS): Dynamically hides specific rows of data (e.g., hiding French sales from the German manager).
2. Object-Level Security (OLS): Hides entire columns or tables (e.g., hiding the entire 'Salary' column from everyone except HR).

How to implement RLS:

1. In Power BI Desktop, go to the Modeling tab and click Manage Roles.
2. Create a Role (e.g., "Germany Manager").
3. Apply a DAX filter to the Geography table: [Country] = "Germany". (This is known as Static RLS).
4. Click View as Role to test it locally. You will see the charts instantly shrink to only show Germany's data!
5. Publish the report to the Power BI Service.
6. In the Service, go to the dataset Security settings, and assign the actual employee's email address to the "Germany Manager" role.

🔄 Data Refresh & The Gateway

Because you used Import Mode (remember Chapter 2?), your cloud report is a frozen snapshot in time. You must tell Power BI to refresh it.

• Manual/On-Demand Refresh: You can click the "Refresh" icon in the dataset menu to pull new data instantly.
• Scheduled Refresh: Go to the dataset settings in the cloud, enter your Data-source credentials so Power BI has permission to read the database, and configure it to refresh daily.
• Dataset Optimization: Ensure you removed unused columns to speed up refresh times! You can check the Refresh History to troubleshoot Refresh failures.
• The On-Premises Data Gateway: If your data lives on a secure, private company server (like an on-prem SQL database), the cloud cannot reach it due to firewalls. You must install a piece of software called a Gateway on your company server. It acts as a secure bridge, allowing the Power BI cloud to reach down into your private network to grab the daily data refresh!

🚨 Alerts and Monitoring

Executives don't want to check the dashboard every single day. They want the dashboard to tell them if something is wrong.

• You can click the ellipsis on a KPI or Card visual on a Dashboard and set a Data Alert by defining Alert thresholds (e.g., "Notify me if Sales < $50k"). Avoid setting unnecessary alerts to prevent alert fatigue!
• Example: "If Total Sales drops below $50,000, send me an email immediately!"
• You can even integrate this with Power Automate to trigger a workflow (e.g., automatically posting a message in Microsoft Teams if a factory machine goes offline).

🚀 Deployment Pipelines

For massive enterprise companies, you don't just publish a report directly to the CEO. You use Deployment Pipelines.

1. Development (Dev): Where the developers build and break things.
2. Testing (Test): Where you publish the beta version. A small group of users click around to find bugs.
3. Production (Prod): The final, perfect version that the whole company sees.

Power BI allows you to migrate a report through these stages. You configure Deployment Rules (e.g., automatically swapping the database connection from the 'Dev SQL Server' to the 'Prod SQL Server' during migration), ensuring strict quality control and controlled releases.

🏋️ Practice Drill

Let's test Row Level Security!

Task:

// Try answering these:
1. In Power BI Desktop, go to the **Modeling** ribbon.
2. Click **Manage Roles**.
3. Click "Create" to make a new role, name it `Test Role`.
4. Click the three dots next to any text table you have (e.g., a table with Countries, Categories, or Names), and select "Add Filter".
5. Set the DAX formula to equal a specific word that exists in your data (e.g., `[Category] = "Clothing"`). Save it.
6. Now click **View as**, check the box for your `Test Role`, and click OK.

← ☁️ Power BI Service and Collaboration | Next Topic → 💻 End-to-End Mini Project