Supported databases › MySQL RDS
MySQL RDS performance investigation with MCP workflows
Amazon RDS for MySQL layers managed infrastructure on top of MySQL, which simplifies patching and backups but adds RDS-specific configuration constraints and monitoring surfaces. DatabaseMCP covers both MySQL-level diagnostics and RDS infrastructure metrics in a single investigation workflow.
Common MySQL RDS performance problems
- Instance class CPU and memory limits: MySQL RDS performance is bounded by the instance class. CPU saturation from inefficient queries and memory pressure from buffer pool overflow are common when instance class is undersized for the actual workload volume.
- EBS volume IOPS limits: gp2 volumes have IOPS limits tied to volume size. Write-heavy workloads hitting these limits show up as high WriteLatency and DiskQueueDepth CloudWatch metrics. Upgrading to gp3 or provisioned IOPS often resolves storage-bound latency spikes.
- Parameter group defaults for production workloads: RDS parameter groups default to conservative values for general workloads. Settings like innodb_buffer_pool_size, max_connections, slow_query_log, and long_query_time often need tuning for production database profiles.
- Read replica lag on write-heavy primaries: Replicas using binlog replication apply writes serially by default. High single-threaded write throughput on the primary causes Replica_SQL_Running_Behind_Master to grow, serving stale reads from replicas.
- Maintenance window impact: RDS automated backups and multi-AZ failovers can cause brief performance degradation or connection interruption. Applications without retry logic surface these as errors.
- Core MySQL issues: Slow queries, InnoDB lock contention, buffer pool pressure, and connection limit saturation all apply to MySQL RDS in the same way as self-hosted MySQL.
MCP investigation workflow for MySQL RDS
- RDS CloudWatch baseline: Start with CPUUtilization, FreeableMemory, ReadIOPS, WriteIOPS, DatabaseConnections, and ReplicaLag to identify which resource layer is under pressure before investigating query-level issues.
- Performance Schema query analysis: Identify top queries by total latency, execution count, and rows examined. Correlate high-cost queries with CloudWatch spikes.
- Slow query log review: Enable the slow query log via the RDS parameter group and review entries for full table scans, high lock wait time, and repeated long-running patterns.
- InnoDB and connection health: Review InnoDB buffer pool hit rate, pending I/O, lock waits, and current connection count to identify engine-level pressure.
- Replication lag monitoring: Track ReplicaLag metric for read replica deployments. Investigate primary write patterns when lag grows consistently.
MySQL RDS and Aurora
For Aurora-specific MySQL deployments seeMySQL Aurora performance investigation. For self-hosted MySQL see MySQL performance investigation.
Get started
See platform features on minidba.com