Description

• Catch subprocess.TimeoutExpired during SAM teardown and force-kill, so slow shutdown cannot fail the entire test suite.
• Added temporary flake-check workflow that runs each test aws-lambda test 5 times to verify this removes the flake (workflow is removed in the last commit; was just to verify that it works). [https://github.com/getsentry/sentry-python/actions/runs/30539137588/job/90862136512?pr=6937](last successful run)
• Switched SAM --warm-containers startup from EAGER to LAZY so Docker containers start on first invocation instead of all 21 at SAM startup. Verified locally by running TESTPATH=tests/integrations/aws_lambda/test_aws_lambda.py uv run tox -e py3.8-aws_lambda both for EAGER & LAZY 10x independently:

EAGER:
=================================================
Metric              |   Mean (s) |  Std Dev (s)
=================================================
Setup time          |      43.69 |         3.99
First call time     |       1.92 |         0.22
Max Teardown time   |       7.49 |         0.27
Total run time      |     108.53 |         4.56
=================================================
LAZY:
=================================================
Metric              |   Mean (s) |  Std Dev (s)
=================================================
Setup time          |      33.32 |         1.82
First call time     |       6.32 |         0.63
Max Teardown time   |       6.90 |         0.18
Total run time      |     102.06 |         3.60
=================================================

Issues

Resolves: #6911 & PY-2632

Reminders

• Please add tests to validate your changes, and lint your code using uv run ruff.
• Add GH Issue ID & Linear ID (if applicable)
• PR title should use conventional commit style (feat:, fix:, ref:, meta:)
• For external contributors: CONTRIBUTING.md, Sentry SDK development docs, Discord community