I evaluate monitoring tools by collection cost, historical context and usefulness in choosing a correction. Customer discussions shaped those three criteria.
- Collection footprint: measure the monitoring workload’s CPU, I/O, memory and query overhead under representative conditions, including periods of pressure.
- Past and present context: retain enough plans, waits, errors and trends to explain a change. Capacity forecasts can help planning, but a forecast is an estimate rather than a guarantee.
- Actionable evidence: require a recommendation to identify its evidence, affected workload and a way to validate or reverse the change. Counters, DMVs and waits can complement one another.
My historical shortlist included Database Performance Analyzer, SQL Diagnostic Manager, SQL Monitor, Spotlight Cloud and SQL Plan Warnings. The order was not a ranking. Those links describe the original recommendations rather than current editions, prices or availability.
Test a candidate against a known incident and the same operational requirements. Check useful before/after context and support for the investigating team. An AI label or many counters does not establish diagnostic quality.
Related reading
- Database Performance Analyzer
- SQL Server Monitoring Week – SQL Diagnostic Manager
- SQL Monitor
- Spotlight Cloud
- SQL Server Monitoring Week – SQL Plan Warnings
An impressive counter inventory is not diagnostic quality, it is useful only when it helps explain and resolve incidents.
Published by Pinal Dave on SQLAuthority. More of my work at pinaldave.com.
Discover more from SQL Authority with Pinal Dave
Subscribe to get the latest posts sent to your email.
