Supported Databases › PostgreSQL Aurora
PostgreSQL Aurora Performance Investigation with MCP Workflows
Amazon Aurora PostgreSQL uses a distributed storage architecture that changes some performance characteristics compared to standard RDS PostgreSQL. DatabaseMCP covers Aurora-specific diagnostics alongside the core PostgreSQL investigation workflow, giving teams a complete picture during incidents.
Common PostgreSQL Aurora performance problems
- Reader node replica lag: Aurora uses a shared storage layer that typically keeps reader nodes very close to the writer, but under heavy write loads or DDL operations the replica lag metric (AuroraReplicaLag) can grow enough to serve stale data for time-sensitive reads.
- Query latency spikes after failover: Automatic failover promotes a reader to writer but can cause a temporary plan cache reset and connection redistribution. Query performance during the minutes after a failover often differs from steady-state behaviour.
- Aurora Serverless v2 scaling delays: Serverless v2 scales ACU capacity in response to load, but scaling is not instantaneous. Sudden workload spikes can hit the current capacity ceiling before the instance scales up, creating a latency window.
- Storage volume I/O pressure: Aurora's distributed storage is durable and auto-scaling, but write-intensive workloads with large transactions or bulk operations can generate high VolumeWriteIOPs that affect writer throughput.
- Connection routing misconfiguration: Applications routing write traffic to reader endpoints or read traffic to the writer endpoint create unnecessary load imbalance. Aurora's reader and writer DNS endpoints should be explicitly separated in application connection strings.
- Core PostgreSQL issues: Slow queries, lock contention, autovacuum lag, and connection pressure all apply to Aurora PostgreSQL the same as any PostgreSQL deployment.
MCP investigation workflow for PostgreSQL Aurora
- Aurora-specific CloudWatch metrics: AuroraReplicaLag, ServerlessDatabaseCapacity, VolumeWriteIOPs, and BufferCacheHitRatio give a fast starting picture of Aurora-layer health.
- Performance Insights: Aurora Performance Insights provides a detailed view of database load broken down by wait events, SQL statements, hosts, and users. It is the primary tool for identifying what is driving load on the Aurora instance.
- Query-level analysis via pg_stat_statements: Identify top queries by total execution time and call count to separate high-frequency low-cost queries from low-frequency high-cost ones.
- Failover and scaling event correlation: Match latency spikes against Aurora events (failovers, scaling events, maintenance windows) to rule out infrastructure causes before investigating query or schema issues.
Related PostgreSQL coverage
For standard RDS deployments see PostgreSQL RDS investigation. For self-hosted PostgreSQL see PostgreSQL investigation.
Get started
See platform features on minidba.com