Database Technology

The hot_standby_feedback GUC allows standbys to inform the primary about running queries, preventing transaction ID wraparound cleanup from cancelling standby work. However, this mechanism specifically addresses cleanup conflicts only; it does not resolve issues arising from schema changes like lock or drop conflicts. Enabling this setting effectively shifts the cost of standby query longevity into increased bloat on the primary database.

• Enabling hot_standby_feedback prevents standby query cancellations due to vacuum cleanup.
• This setting trades primary disk bloat for improved standby read availability.
• It does not mitigate conflicts caused by schema drops or locking changes.
• Monitor primary bloat levels closely if this GUC is enabled on large fleets.

AWS introduces practical guidance for building applications on the new Aurora DSQL distributed architecture. The blog outlines how to identify scalability bottlenecks and apply proven patterns for workload distribution. Key topics include primary key selection, schema design, indexing strategies, and multi-Region optimization while maintaining full ACID compliance.

• Focus on primary key selection and schema design to prevent data skew in distributed shards.
• Implement indexing strategies that align with Aurora DSQL's distributed query execution engine.
• Optimize transaction strategies to maintain ACID compliance across multiple AWS Regions.
• Use multi-Region deployment patterns to reduce latency and improve availability for global apps.

A new AWS Database Blog post outlines a structured workflow for diagnosing T-SQL performance issues on Amazon RDS for SQL Server. It details how to integrate CloudWatch Database Insights, Query Store, and Resource Governor to detect plan regressions and isolate heavy analytical workloads. This approach helps administrators systematically identify and resolve execution plan changes that degrade query speed.

• Combine CloudWatch Database Insights with Query Store for end-to-end visibility into query regressions.
• Use Resource Governor to isolate and throttle analytical workloads impacting transactional performance.
• Follow a standardized workflow to detect, diagnose, and resolve T-SQL execution plan changes.
• Leverage built-in RDS tools to identify specific queries causing resource contention.

Fleet impact: For DBAs managing RDS SQL Server, this workflow provides a standardized method to catch execution plan regressions before they impact SLAs. Ensure Query Store is enabled and configured to capture historical data, and review Resource Governor configurations to effectively segment analytical versus OLTP workloads.

AWS now enables a serverless pipeline to irreversibly redact personally identifiable information from Amazon RDS for PostgreSQL audit logs. The system detects and masks over 30 data types, including SSNs, credit cards, and names, before archiving the clean logs to Amazon S3. Once stored, the redacted logs remain fully queryable via Amazon Athena for compliance and forensic analysis.

• Automated removal of SSNs, credit cards, emails, and 30+ other PII types from audit streams.
• Serverless pipeline ensures irreversible redaction before data reaches Amazon S3 storage.
• Archived logs stay queryable via Amazon Athena for ongoing compliance and auditing needs.