chore(ci_visibility): ensure exception dictionary is cleared after each test [backport 3.12] (#14346)

Backport a5aee01 from #14344 to 3.12.

The solution in PR #14328 was incomplete. This PR ensures all exceptions
are cleared up at the end of each test.

I'm not adding a release note because this is a continuation of #14328,
which already has a release note.

## Checklist
- [x] PR author has checked that all the criteria below are met
- The PR description includes an overview of the change
- The PR description articulates the motivation for the change
- The change includes tests OR the PR description describes a testing
strategy
- The PR description notes risks associated with the change, if any
- Newly-added code is easy to change
- The change follows the [library release note
guidelines](https://ddtrace.readthedocs.io/en/stable/releasenotes.html)
- The change includes or references documentation updates if necessary
- Backport labels are set (if
[applicable](https://ddtrace.readthedocs.io/en/latest/contributing.html#backporting))

## Reviewer Checklist
- [x] Reviewer has checked that all the criteria below are met 
- Title is accurate
- All changes are related to the pull request's stated goal
- Avoids breaking
[API](https://ddtrace.readthedocs.io/en/stable/versioning.html#interfaces)
changes
- Testing strategy adequately addresses listed risks
- Newly-added code is easy to change
- Release note makes sense to a user of the library
- If necessary, author has acknowledged and discussed the performance
implications of this PR as reported in the benchmarks PR comment
- Backport labels are set in a manner that is consistent with the
[release branch maintenance
policy](https://ddtrace.readthedocs.io/en/latest/contributing.html#backporting)

Co-authored-by: Vítor De Araújo <vitor.dearaujo@datadoghq.com>

Author: dd-octo-sts[bot]

ddtrace/contrib/internal/pytest/_plugin_v2.py

@@ -581,16 +581,21 @@ def _pytest_runtest_protocol_post_yield(item, nextitem, coverage_collector):
     if coverage_collector is not None and not InternalTest.is_finished(test_id):
         _handle_collected_coverage(item, test_id, coverage_collector)
 
+    reports_dict = reports_by_item.pop(item, None)
+
     if not InternalTest.is_finished(test_id):
         log.debug("Test %s was not finished normally during pytest_runtest_protocol, finishing it now", test_id)
-        reports_dict = reports_by_item.pop(item, None)
         if reports_dict:
             test_outcome = _process_reports_dict(item, reports_dict)
             InternalTest.finish(test_id, test_outcome.status, test_outcome.skip_reason, test_outcome.exc_info)
         else:
             log.debug("Test %s has no entry in reports_by_item", test_id)
             InternalTest.finish(test_id)
 
+    if reports_dict:
+        for report in reports_dict.values():
+            excinfo_by_report.pop(report, None)
+
     # We rely on the CI Visibility service to prevent finishing items that have been discovered and have unfinished
     # children, but as an optimization:
     # - we know we don't need to finish the suite if the next item is in the same suite

tests/contrib/pytest/test_pytest.py

@@ -12,6 +12,8 @@
 from ddtrace.contrib.internal.pytest._utils import _pytest_version_supports_atr
 from ddtrace.contrib.internal.pytest._utils import _pytest_version_supports_efd
 from ddtrace.contrib.internal.pytest._utils import _pytest_version_supports_itr
+from ddtrace.contrib.internal.pytest._utils import excinfo_by_report
+from ddtrace.contrib.internal.pytest._utils import reports_by_item
 from ddtrace.contrib.internal.pytest.constants import XFAIL_REASON
 from ddtrace.contrib.internal.pytest.patch import get_version
 from ddtrace.contrib.internal.pytest.plugin import is_enabled
@@ -4591,6 +4593,21 @@ def test_log_capture():
         assert "I/O operation on closed file" not in result.stderr.str()
         assert result.ret == 0
 
+    def test_pytest_clears_excinfo_dict_after_use(self):
+        reports_by_item_count_before = len(reports_by_item)
+        excinfo_by_report_count_before = len(excinfo_by_report)
+
+        self.testdir.makepyfile(
+            """
+            def test_one():
+                assert False
+            """
+        )
+
+        self.inline_run("--ddtrace")
+        assert len(reports_by_item) == reports_by_item_count_before
+        assert len(excinfo_by_report) == excinfo_by_report_count_before
+
     @pytest.mark.skipif(
         not _PYTEST_SUPPORTS_ITR or not _PYTEST_SUPPORTS_EFD,
         reason=f"pytest version {get_version()} does not support EFD or ITR coverage reporting",