Feature Suggestions

[rkhouani-db]

Direct Query History Access:
Add an option to fetch queries directly from the query history without requiring them to be stored in a folder first.

Query Execution Randomization:
Introduce an option to randomize the order of query execution. This could be particularly useful for testing and simulating real-world scenarios with varying query patterns.

Dynamic Concurrency Simulation:
Enable dynamic concurrency by calculating the average queries per minute from the query history. This feature would allow users to simulate a full day of production usage more realistically, adapting to historical query patterns.

[TravBricks]

Query history capture would be really awesome, but it could be two mechanisms.

Mechanism 1: Harvest query history
Harvest query history to sql files that are named with the sequence
Stretch goal: Augment queries with a transaction version, so data is static for all benchmarks

Mechanism 2: Replay
Have metimur run benchmarks that can run through the sequence as a replay
Stretch goal: Specific start and end time within the replay for quick testing