Performance & health review
Find out what your database is actually doing.
A structured review of a production MySQL estate, sized in hours to the size of the system. The output is not a report that restates dashboards — it is a ranked fix list: each finding with its mechanism, its expected impact, and the concrete change that addresses it.
What we examine
- Workload: slow-query log and Performance Schema digests — which queries dominate load, which are about to, and what their plans actually do.
- Indexes and schema: missing, redundant, and unused indexes; datatype and collation problems; primary-key design and its effect on secondary indexes.
- Server and InnoDB configuration: buffer pool sizing and hit rates, redo and flush settings, connection handling — tuned against your workload, not a generic checklist.
- Replication and backup posture: topology, lag behavior under load, whether backups restore (we check, because "we have backups" and "we can restore" are different claims), and how far you actually are from your stated recovery objectives.
How it runs
The review works from read access: slow logs, Performance Schema, schema dumps, and monitoring. We do not change production during the review phase. Findings are walked through with your engineers — every item explained to the level where your team could defend the change without us in the room.
What you leave with
A prioritized list where the top items are typically worth the whole engagement: the covering index that removes a filesort from your hottest endpoint, the configuration line silently halving your buffer pool, the table whose missing primary key explains your replica lag. Where you want help executing, the same engineers carry the fixes through.