Luis Núñez

In this post, I summarize my hands-on experience completing the Microsoft Applied Skills: Configure and migrate to Azure Database for PostgreSQL assessment.

Using Azure Portal, pgAdmin, and PostgreSQL, I walked through the core architectural tasks of setting up, securing, migrating, replicating, and optimizing a managed relational database in the cloud.

🛠️ Step-by-Step Lab Execution

1. Security & Monitoring Setup

  • Authentication: Configured dual authentication on Azure Database for PostgreSQL Flexible Server (psql63693133), enabling Microsoft Entra ID alongside native PostgreSQL authentication.
  • Network Isolation: Isolated server traffic within VNET1/subnet1 by creating a Private Endpoint (psql63693133-pe).
  • Diagnostics: Configured Diagnostic Settings (logs-to-loganalytics) to route PostgreSQL Sessions data logs to an Azure Log Analytics Workspace (loganalytics1).

2. Migration & Handling the B-Tree Index Error

  • Data Restore: Restored the adventureworks sample database to Azure using pgAdmin's restore tool.
  • Real-World Error: During pg_restore, a known error occurred due to the PostgreSQL B-tree index row size limit: 
ERROR: index row size 2896 exceeds btree version 4 maximum 2704 for index "IX_ProductReview_ProductID_Name"

Enter fullscreen mode Exit fullscreen mode

  • Resolution & Impact: This error was non-fatal. Because PostgreSQL restores table rows before creating secondary indexes, all schema structures and data were successfully imported. Only the secondary index on productreview.comments failed due to text size limits.

3. Cross-Region Read Replication

  • High Availability: Created a Read Replica (psql63693133-replica) to offload read-heavy workloads and increase resiliency.
  • Paired Region: Deployed the replica in Canada Central (the paired region for the primary server located in Canada East).

4. Query Performance Optimization

  • EXPLAIN Plan: Analyzed the public.address1() function in pgAdmin, extracted its underlying SQL query, generated the EXPLAIN execution plan, and exported the visual output.
  • Index Creation: Identified a sequential scan bottleneck on person.address and resolved it by creating a targeted index: 
CREATE INDEX IX_Address_StateProvinceID 
ON person.address (stateprovinceid);

Enter fullscreen mode Exit fullscreen mode

💡 Key Takeaways

  • Private Endpoints First: Always restrict database traffic to private virtual networks before migrating production data.
  • Understand Migration Logs: Non-critical index failures (like B-tree row limits on text columns) do not invalidate full data restores.
  • Analyze Before Scaling: Using EXPLAIN to spot missing indexes solves query bottlenecks without needing to increase expensive server compute tiers.

🚀 Try It Yourself

Put your hands-on cloud skills to the test with the official assessment module on Microsoft Learn:

👉 Microsoft Applied Skills: Configure and migrate to Azure Database for PostgreSQL