query-performance-tuning

Diagnose slow database queries by capturing an actual execution plan with realistic parameters, locating the dominant slow node, and classifying the root cause across wrong plans, wrong access paths, or working-set issues to propose validated remediation steps.

Denormalization becomes warranted over indexing when query tuning rules indicate that computed columns, materialized views, or read stores better resolve the slow database query's working-set limits or access path inefficiencies than standard composite or covering indexes.

Detect N+1 query patterns by analyzing slow-query logs or traces to identify repeated sequential access paths, then fix them through query rewrites or targeted index selection to optimize overall database performance and reduce redundant sequential scans.

Resolve disk-bound thrashing and working-set issues by sizing memory requirements, then selecting capacity interventions like memory scaling, data partitioning, archiving, or index bloat recovery to ensure the active working set fits efficiently in memory.

Yes, query tuning applies to both PostgreSQL and MySQL by interpreting execution plan signals such as sequential versus index scans, row estimates versus actuals, filter effectiveness, and sort or hash join behavior to pinpoint the slow node.

Indexes may be mis-chosen due to outdated database statistics, leading to wrong access paths or incorrect row estimates in the execution plan; resolving this requires stats updates, query rewrites, or index design adjustments like partial or covering index patterns.