[Varshith-Kali]

Fixes #4853

Root cause

scripts/update_schemas_from_aws_api.py built the per-engine class data with dict.fromkeys(available_results.keys(), dict()), which hands every engine the same dict object. Every minor-version bucket then accumulated the union of all engines' orderable classes, so all 14 engines carried identical DBInstanceClass enums and branches for versions they do not even have (e.g. aurora-mysql had a branch for the MariaDB version 10.11). E3062 passed invalid engine/class/version combos, like mysql 5.7 on db.t4g.medium, which then fails at deploy time with CREATE_FAILED. This aliasing has been in the script since E3062 was added, so the data file was never correct.

Changes

• scripts/update_schemas_from_aws_api.py: build a fresh dict per engine instead of sharing one via dict.fromkeys. Future regenerations (which need AWS credentials) will now produce correct per-engine data.
• src/cfnlint/data/schemas/extensions/aws_rds_dbinstance/db_instance_class.json:
  • Dropped 842 branches that pair an engine with a minor version that engine does not have, keeping 138. Validity was checked against engine_version.json from the same data run (e.g. aurora-mysql now only has 5.7/8.0/8.4, mariadb only 10.5/10.6/10.11/11.4/11.8/12.3).
  • Scoped the mysql 5.7 enum to the 22 classes in the db.m5/db.r5/db.t3 families, which is what RDS offers for mysql 5.7.44 in us-east-1 per the issue (the docs also give db.t4g.medium on MySQL as 8.4 and 8.0). All currently available mysql 5.7 versions are 5.7.44-rds.* builds, so this covers the whole branch without false-positive risk.
• test/unit/rules/resources/rds/test_db_instance_class_data.py: new regression tests.

Test evidence

• New tests (5) written against the old data first: 3 failed as expected (repro template passed with no findings, mysql 5.7 enum contained db.t4g.medium, 842 cross-engine branches present), 2 passed.
• After the fix: all 5 pass, and all 134 tests in test/unit/rules/resources/rds/ pass.
• The issue's exact repro template (mysql, 5.7.44-rds.20260902, db.t4g.medium) now produces a single E3062 error; db.r5.large on the same engine/version and db.t4g.medium on mysql 8.0 still pass.
• ruff check and ruff format --check clean on changed files. The JSON was rewritten with the generator's exact serialization settings, and a round-trip check on the original file confirmed byte stability, so the diff is purely the data correction.

Note

For minor versions shared by engines in the same family (e.g. mysql and aurora-mysql on 8.0, sqlserver flavors on 15.00), the enums remain the union across those engines, since I could not source the exact per-engine split without AWS API access. The script fix means the next regeneration with credentials will produce the exact per-engine enums. Nothing in this PR makes any currently passing valid combo fail: branch pruning only removed engine/version pairings the engine does not have, and the mysql 5.7 scoping matches the authoritative data in the issue.

[Varshith-Kali]

Hi maintainers, the CI workflows on this PR are waiting on the first-time-contributor approval gate. Would appreciate an approval to run them when you get a moment. Happy to address any review feedback. Thanks!

[Varshith-Kali]

Hi team, on this fix(rds) PR: the [CI] Test PR workflow run on the current head is still parked waiting for a first-time-contributor approval (action_required). Could a maintainer approve the run so CI can proceed? Thanks.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.69%. Comparing base (8ec570a) to head (83df052).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #4870   +/-   ##
=======================================
  Coverage   94.69%   94.69%           
=======================================
  Files         442      442           
  Lines       16083    16083           
  Branches     3119     3119           
=======================================
  Hits        15229    15229           
  Misses        470      470           
  Partials      384      384           

Flag	Coverage Δ
unittests	94.68% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.