How to Confirm the Effect of Response Improvement in Custom Measures

We are considering how to confirm the effect after performance improvements of custom measures.
Currently, we check by measuring the elapsed time until the table opens,
but this method has the following issues.

Example:
Before improvement: 10 minutes
After improvement: 7 minutes
Timeout setting: 5 minutes

In this case, even after the improvement, it will still time out, so the effect
cannot be confirmed even if performance has actually improved.

Therefore,

• Effective diagnostic tools and logs
• Methods to check query execution time and processing costs
• Ways to evaluate the load per custom measure
• Performance evaluation methods recommended by Oracle

We would appreciate advice from anyone who knows performance evaluation methods