Database Technology

AWS has introduced an assistant to help configure connection pooling for the AWS Advanced JDBC Wrapper on Amazon Aurora and RDS. The guide clarifies the differences between internal and external pooling mechanisms to aid in selection. This tool assists engineers in generating the correct configuration parameters for their specific database workloads.

• Use the JDBC Wrapper Configuration Assistant to automate pool setup
• Distinguish between internal and external pooling strategies
• Optimize connection handling for Aurora and RDS workloads
• Reduce manual configuration errors in Java application stacks

Wellingtone Luvonga details the complexities of multi-region PostgreSQL disaster recovery using Crunchy PGO. The guide addresses specific hurdles like enforcing secure TLS endpoints for pgBackRest and avoiding timeline conflicts during failback. It provides a full lifecycle walkthrough from secure bootstrap to handling S3 archive integrity.

• Configuring native SSL for local storage or MinIO is often an administrative burden in HA setups.
• Multi-region DR introduces risks like timeline conflicts and S3 archive poisoning during failback.
• Crunchy PGO provides a structured approach to managing secure replication and failover.
• The guide covers the complete lifecycle, including safe failback procedures to the original primary.
• Secure TLS endpoints are critical for pgBackRest operations in distributed architectures.

AWS has introduced a revert-export command within the DynamoDB Bulk Executor tool to undo accidental table modifications. By leveraging incremental exports to Amazon S3, operators can selectively restore data without performing a full table restore. The tool supports filtering specific changes via transforms or correcting individual items to minimize recovery time and data loss.

• Use revert-export to undo unwanted writes without full table restores
• Leverage incremental S3 exports for efficient, targeted rollback operations
• Apply transforms to fix subsets of changes or specific items
• Avoids the latency and cost of restoring entire DynamoDB tables

AWS outlines a pattern to maintain write availability in Amazon Neptune by decoupling write acceptance from execution using message queues like SQS, Kinesis, or MSK. This architecture ensures applications continue accepting graph writes during maintenance windows, failovers, and scaling events. The approach isolates the database from transient write bursts or interruptions by buffering requests in a queue.

• Use SQS, Kinesis, or MSK to buffer writes, decoupling acceptance from execution.
• Maintains write availability during Neptune maintenance, failovers, and scaling.
• Prevents write failures when the graph engine is temporarily unavailable.
• Requires application logic to handle async write acknowledgment and retries.

Stripe engineers have automated database incident recovery by modeling their global infrastructure as a graph. They employ graph search algorithms alongside state machines to compute and execute remediation plans without manual intervention. This approach allows for systematic identification and resolution of database issues across their distributed systems.

• Modeling infrastructure as a graph enables precise dependency mapping for automated recovery.
• Graph search algorithms help identify optimal remediation paths in complex topologies.
• State machines ensure deterministic execution of automated database repair actions.
• Reduces mean time to recovery by removing manual intervention from incident response.
• Demonstrates practical application of graph theory in large-scale database operations.