FIX: Connection pooling flaky tests

[bewithgaurav]

Work Item / Issue Reference

AB#38075

GitHub Issue: #<ISSUE_NUMBER>

---

Summary

This pull request enhances the connection pooling tests in tests/test_003_connection.py by introducing more detailed and robust test cases. The changes include adding new tests for connection pooling performance and connection reuse, as well as improving the structure and clarity of the existing test logic.

Enhancements to connection pooling tests:

• Performance testing for pooled vs. non-pooled connections:

  • Added a new test (test_connection_pooling_speed) to measure and compare the average time taken for establishing connections with and without pooling. The test ensures that pooled connections are at least 20% faster than non-pooled connections.

• Connection reuse validation:

  • Introduced a new test (test_connection_pooling_reuse_spid) to verify that connections are reused from the pool. The test checks that the SQL Server process ID (SPID) remains consistent across reused connections, indicating proper pooling behavior.

• Improved test setup and cleanup:

  • Added explicit enabling and disabling of connection pooling at the start and end of tests to ensure isolation and prevent interference with other tests.

These changes improve the reliability and comprehensiveness of the connection pooling tests.

[Copilot]

Pull Request Overview

This PR addresses flaky connection pooling tests by improving test reliability and adding more comprehensive validation. The changes focus on making the connection pooling tests more robust and deterministic.

• Enhanced connection speed comparison tests with statistical averaging and warm-up procedures
• Added new test to verify actual connection reuse through SQL Server Process ID (SPID) validation
• Improved test isolation with explicit pooling state management

[sumitmsft]

Added couple of comments. Please validate