What problem does it solve? PostgreSQL deployments degrade under real load through replication lag, table bloat, connection exhaustion, lock contention, and untested failover plans. This Skill provides engine-level administration guidance grounded in how PostgreSQL actually behaves, replacing generic tuning folklore with workload-specific recommendations that include rollback paths. ## Core Features & Use Cases - Configuration and Memory Tuning: Recommends shared_buffers, work_mem, effective_cache_size, checkpoint, and planner cost settings matched to actual hardware and workload type. - Vacuum, Replication, and Locking Analysis: Diagnoses autovacuum falling behind, designs streaming or logical replication topologies, and explains lock modes and deadlock patterns for write-heavy tables. - Safe Migration and Recovery Planning: Plans lock-minimizing schema changes on large live tables, partitioning conversions, and backup/PITR strategies with stated downtime and rollback steps. - Use Case: A production database shows growing replication lag and connection pool exhaustion. The Skill identifies whether the root cause is vacuum lag, lock contention, or pooler misconfiguration, then recommends the smallest safe change with expected impact and residual risk. ## Quick Start Ask the agent to diagnose why your PostgreSQL database shows replication lag and connection exhaustion, and recommend a safe fix with a rollback plan.
