Skip to main content

Tuning targets

DBtune allows you to select a tuning target to focus optimization on specific performance goals.

Workload Fingerprint​

Workload Fingerprint lets you optimize for a specific set of queries instead of all queries in your database. This provides the most targeted optimization and is the default tuning target.

Workload Fingerprint automatically analyzes queries based on average query runtime (AQR) and query call count. You can either use DBtune's automatically recommended fingerprints or create your own custom fingerprints by selecting specific queries.

When to use:

  • Optimizing specific workload types (OLTP, OLAP, or mixed)
  • Focusing on critical queries that need performance improvement
  • More granular control over which queries to optimize

Learn more about Workload Fingerprints.

Average query runtime (AQR)​

Average query runtime measures latency in milliseconds (ms).

DBtune computes AQR using total_exec_time / calls from the pg_stat_statements table.

When to use:

  • General-purpose optimization
  • Reducing latency for complex queries
  • Systems not hitting hard resource bottlenecks

Transactions per second (TPS)​

Transactions per second measures throughput. DBtune computes this from the xact_commit metric in the pg_stat_database table.

When to use:

  • System hitting resource bottleneck (close to 100% CPU or I/O)
  • Maximizing throughput at peak load
Last updated on