FIX: Credential instance cache

[jahnvi480]

Work Item / Issue Reference

AB#43254

GitHub Issue: #388

---

Summary

This pull request introduces credential instance caching for Azure Active Directory (AAD) authentication in the mssql_python package, improving performance by reusing credential objects instead of creating new ones for each token request. It also adds comprehensive tests for the caching logic and edge cases, and includes a new benchmark script to measure the performance impact of credential caching.

Credential instance caching and logic changes:

• Added a module-level credential cache (_credential_cache) and lock in auth.py to reuse Azure Identity credential instances per authentication type, enabling the SDK's in-memory token cache and reducing redundant token acquisitions. (mssql_python/auth.py)
• Modified AADAuth.get_raw_token to use the cached credential instance rather than creating a new one each call, ensuring efficient token reuse. (mssql_python/auth.py )

Testing improvements:

• Added a pytest fixture to clear the credential cache between tests, ensuring test isolation. (tests/test_008_auth.py )
• Introduced a new TestCredentialInstanceCache class with tests verifying credential reuse, separation by auth type, and cache state. (tests/test_008_auth.py)
• Added tests for error handling and edge cases, including missing Azure libraries, authentication errors, and unusual connection string parameter cases. (tests/test_008_auth.py )

Benchmarking:

• Added a new benchmark script (benchmarks/bench_credential_cache.py) to measure and compare the performance of credential caching versus the previous behavior of creating new credentials for every token request. (benchmarks/bench_credential_cache.py)

[Copilot]

Pull request overview

This PR adds module-level caching of Azure Identity credential instances in mssql_python AAD auth flows to reduce repeated credential construction and enable Azure Identity’s in-memory token caching, along with new tests and a benchmark for measuring the impact.

Changes:

• Added a module-level credential instance cache (with locking) used by AADAuth._acquire_token() / get_raw_token().
• Expanded auth test coverage to validate credential reuse and additional error/edge paths.
• Added a benchmark script to compare “new credential per call” vs “cached credential instance” token acquisition.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
mssql_python/auth.py	Introduces a global credential-instance cache guarded by a lock and updates token acquisition to reuse credentials.
tests/test_008_auth.py	Adds cache-clearing fixture and new tests for credential reuse, import-error handling, and edge cases.
benchmarks/bench_credential_cache.py	Adds a standalone benchmark script to quantify the performance effect of credential-instance caching.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

📊 Code Coverage Report

🔥 Diff Coverage 100%

🎯 Overall Coverage 77%

📈 Total Lines Covered: 5680 out of 7341 📁 Project: mssql-python

---

Diff Coverage

Diff: main...HEAD, staged and unstaged changes

• mssql_python/auth.py (100%)

Summary

• Total: 7 lines
• Missing: 0 lines
• Coverage: 100%

---

📋 Files Needing Attention

📉 Files with overall lowest coverage (click to expand)

mssql_python.pybind.logger_bridge.hpp: 58.8%
mssql_python.pybind.logger_bridge.cpp: 59.2%
mssql_python.pybind.ddbc_bindings.h: 67.8%
mssql_python.pybind.ddbc_bindings.cpp: 69.7%
mssql_python.row.py: 70.5%
mssql_python.pybind.connection.connection.cpp: 75.3%
mssql_python.__init__.py: 77.1%
mssql_python.ddbc_bindings.py: 79.6%
mssql_python.pybind.connection.connection_pool.cpp: 79.6%
mssql_python.connection.py: 85.2%

---

🔗 Quick Links

⚙️ Build Summary	📋 Coverage Details
View Azure DevOps Build	Browse Full Coverage Report