[cnv-4.19] Add marker for class/module scoped data collection (#4118)

conftest.py

@@ -33,6 +33,7 @@
 from utilities.data_collector import (
     collect_default_cnv_must_gather_with_vm_gather,
     get_data_collector_dir,
+    get_scope_identifier,
     set_data_collector_directory,
     set_data_collector_values,
 )
@@ -644,12 +645,13 @@ def pytest_runtest_setup(item):
         # before the setup work starts, insert current epoch time into the database
         try:
             db = Database(base_dir=item.config.getoption("--data-collector-output-dir"))
-            db.insert_test_start_time(
-                test_name=f"{item.fspath}::{item.name}",
-                start_time=int(datetime.datetime.now().strftime("%s")),
-            )
+            scope_marker = item.get_closest_marker(name="data_collector_scope")
+            scope_value = scope_marker.kwargs.get("scope") if scope_marker else None
+
+            name = get_scope_identifier(node=item, scope_value=scope_value)
+            db.insert_start_time(name=name, start_time=int(datetime.datetime.now().strftime("%s")))
         except Exception as db_exception:
-            LOGGER.error(f"Database error: {db_exception}. Must-gather collection may not be accurate")
+            LOGGER.error(f"[DATA_COLLECTOR] Database error: {db_exception}. Must-gather collection may not be accurate")
     BASIC_LOGGER.info(f"\n{separator(symbol_='-', val=item.name)}")
     BASIC_LOGGER.info(f"{separator(symbol_='-', val='SETUP')}")
     if "incremental" in item.keywords:
@@ -852,29 +854,28 @@ def get_inspect_command_namespace_string(node: Node, test_name: str) -> str:
 
 def calculate_must_gather_timer(test_start_time):
     if test_start_time > 0:
-        return int(datetime.datetime.now().strftime("%s")) - test_start_time
+        # Add 5-minute (300s) buffer to work around must-gather timing issues
+        return int(datetime.datetime.now().strftime("%s")) - test_start_time + 300
     else:
-        LOGGER.warning(f"Could not get start time of test. Collecting must-gather for last {TIMEOUT_5MIN}s")
+        LOGGER.warning(f"[DATA_COLLECTOR] Could not get start time. Collecting must-gather for last {TIMEOUT_5MIN}s")
         return TIMEOUT_5MIN
 
 
 def pytest_exception_interact(node: Item | Collector, call: CallInfo[Any], report: TestReport | CollectReport) -> None:
     BASIC_LOGGER.error(report.longreprtext)
     if node.config.getoption("--data-collector") and not is_skip_must_gather(node=node):
         test_name = f"{node.fspath}::{node.name}"
-        LOGGER.info(f"Must-gather collection is enabled for {test_name}.")
+        LOGGER.info(f"[DATA_COLLECTOR] Must-gather collection is enabled for {test_name}.")
         inspect_str = get_inspect_command_namespace_string(test_name=test_name, node=node)
         if call.excinfo and any([
             isinstance(call.excinfo.value, exception_type) for exception_type in MUST_GATHER_IGNORE_EXCEPTION_LIST
         ]):
-            LOGGER.warning(f"Must-gather collection would be skipped for exception: {call.excinfo.type}")
+            LOGGER.warning(
+                f"[DATA_COLLECTOR] Must-gather collection would be skipped for exception: {call.excinfo.type}"
+            )
         else:
-            try:
-                db = Database(base_dir=node.config.getoption("--data-collector-output-dir"))
-                test_start_time = db.get_test_start_time(test_name=test_name)
-            except Exception as db_exception:
-                test_start_time = 0
-                LOGGER.warning(f"Error: {db_exception} in accessing database.")
+            db = Database(base_dir=node.config.getoption("--data-collector-output-dir"))
+            test_start_time = db.get_start_time_for_collection(node=node)
 
             try:
                 collection_dir = os.path.join(get_data_collector_dir(), "pytest_exception_interact")

pytest.ini

@@ -15,6 +15,7 @@ markers =
 
     # Data collection markers
     skip_must_gather_collection: skip must gather collection on failures
+    data_collector_scope: Set data collection scope (test, class, or module) for must-gather timing
 
     # Test types
     destructive: Destructive tests

tests/virt/cluster/aaq/test_aaq_allocation_methods.py

@@ -14,6 +14,7 @@
         "updated_namespace_with_aaq_label",
     ),
     pytest.mark.arm64,
+    pytest.mark.data_collector_scope(scope="module"),
 ]
 
 

tests/virt/cluster/aaq/test_arq.py

@@ -39,6 +39,7 @@
         "updated_namespace_with_aaq_label",
     ),
     pytest.mark.gating,
+    pytest.mark.data_collector_scope(scope="module"),
 ]
 
 

tests/virt/cluster/common_templates/centos/test_centos_os_support.py

@@ -25,6 +25,8 @@
     wait_for_console,
 )
 
+pytestmark = pytest.mark.data_collector_scope(scope="module")
+
 LOGGER = logging.getLogger(__name__)
 TESTS_CLASS_NAME = "TestCommonTemplatesCentos"
 

tests/virt/cluster/common_templates/fedora/test_fedora_os_support.py

@@ -28,6 +28,8 @@
     wait_for_console,
 )
 
+pytestmark = pytest.mark.data_collector_scope(scope="module")
+
 LOGGER = logging.getLogger(__name__)
 TESTS_CLASS_NAME = "TestCommonTemplatesFedora"
 

tests/virt/cluster/common_templates/rhel/test_rhel_os_support.py

@@ -30,7 +30,11 @@
     wait_for_console,
 )
 
-pytestmark = [pytest.mark.post_upgrade, pytest.mark.gating]
+pytestmark = [
+    pytest.mark.post_upgrade,
+    pytest.mark.gating,
+    pytest.mark.data_collector_scope(scope="module"),
+]
 
 
 LOGGER = logging.getLogger(__name__)

tests/virt/cluster/common_templates/windows/test_windows_os_support.py

@@ -26,7 +26,12 @@
     validate_virtctl_guest_agent_data_over_time,
 )
 
-pytestmark = [pytest.mark.post_upgrade, pytest.mark.special_infra, pytest.mark.high_resource_vm]
+pytestmark = [
+    pytest.mark.post_upgrade,
+    pytest.mark.special_infra,
+    pytest.mark.high_resource_vm,
+    pytest.mark.data_collector_scope(scope="module"),
+]
 
 LOGGER = logging.getLogger(__name__)
 TESTS_CLASS_NAME = "TestCommonTemplatesWindows"

tests/virt/cluster/migration_and_maintenance/test_evictionstrategy.py

@@ -23,7 +23,11 @@
 
 LOGGER = logging.getLogger(__name__)
 
-pytestmark = [pytest.mark.arm64, pytest.mark.rwx_default_storage]
+pytestmark = [
+    pytest.mark.arm64,
+    pytest.mark.rwx_default_storage,
+    pytest.mark.data_collector_scope(scope="module"),
+]
 
 
 def wait_for_vm_uid_mismatch(vmi, vmi_old_uid):

tests/virt/cluster/migration_and_maintenance/test_migration_policy.py

@@ -11,6 +11,8 @@
     running_vm,
 )
 
+pytestmark = pytest.mark.data_collector_scope(scope="module")
+
 NAMESPACE_LABEL = {"awesome-namespace-label": ""}
 
 DEFAULT_MIGRATION_POLICY_PARAMETERS = {

tests/virt/node/general/test_machinetype.py

@@ -15,7 +15,7 @@
     wait_for_updated_kv_value,
 )
 
-pytestmark = pytest.mark.post_upgrade
+pytestmark = [pytest.mark.post_upgrade, pytest.mark.data_collector_scope(scope="module")]
 LOGGER = logging.getLogger(__name__)
 
 

tests/virt/node/hotplug/test_cpu_memory_hotplug.py

@@ -27,7 +27,7 @@
     restart_vm_wait_for_running_vm,
 )
 
-pytestmark = pytest.mark.rwx_default_storage
+pytestmark = [pytest.mark.rwx_default_storage, pytest.mark.data_collector_scope(scope="module")]
 
 
 LOGGER = logging.getLogger(__name__)

tests/virt/node/migration_and_maintenance/test_node_maintenance.py

@@ -30,7 +30,11 @@
     start_and_fetch_processid_on_windows_vm,
 )
 
-pytestmark = [pytest.mark.post_upgrade, pytest.mark.rwx_default_storage]
+pytestmark = [
+    pytest.mark.post_upgrade,
+    pytest.mark.rwx_default_storage,
+    pytest.mark.data_collector_scope(scope="module"),
+]
 
 
 LOGGER = logging.getLogger(__name__)

tests/virt/node/migration_and_maintenance/test_post_copy_migration.py

@@ -23,7 +23,11 @@
     start_and_fetch_processid_on_windows_vm,
 )
 
-pytestmark = [pytest.mark.rwx_default_storage, pytest.mark.usefixtures("created_post_copy_migration_policy")]
+pytestmark = [
+    pytest.mark.rwx_default_storage,
+    pytest.mark.usefixtures("created_post_copy_migration_policy"),
+    pytest.mark.data_collector_scope(scope="module"),
+]
 
 
 LOGGER = logging.getLogger(__name__)

utilities/data_collector.py

@@ -4,9 +4,11 @@
 import shlex
 from functools import cache
 
+from _pytest.nodes import Collector
 from ocp_resources.namespace import Namespace
 from ocp_resources.resource import get_client
 from ocp_utilities.monitoring import Prometheus
+from pytest import Item
 from pytest_testconfig import config as py_config
 
 import utilities.hco
@@ -175,3 +177,22 @@ def prepare_pytest_item_data_dir(item, output_dir):
     )
     os.makedirs(item_dir_log, exist_ok=True)
     return item_dir_log
+
+
+def get_scope_identifier(node: Item | Collector, scope_value: str | None) -> str:
+    """
+    Get the identifier name based on data collection scope.
+
+    Args:
+        node: Pytest node (Item or Collector).
+        scope_value: Scope value from marker ("module", "class", or None for test).
+
+    Returns:
+        Database key for this scope.
+    """
+    if scope_value == "module":
+        return str(node.fspath)
+    elif scope_value == "class":
+        return f"{node.fspath}::{node.parent.name}" if node.parent else str(node.fspath)
+    else:
+        return f"{node.fspath}::{node.name}"

utilities/database.py

@@ -1,9 +1,12 @@
+import datetime
 import logging
 
+from _pytest.nodes import Collector
+from pytest import Item
 from sqlalchemy import Integer, String, create_engine
 from sqlalchemy.orm import DeclarativeBase, Mapped, Session, mapped_column
 
-from utilities.data_collector import get_data_collector_base
+from utilities.data_collector import get_data_collector_base, get_scope_identifier
 
 LOGGER = logging.getLogger(__name__)
 
@@ -32,17 +35,68 @@ def __init__(
         self.engine = create_engine(url=self.connection_string, echo=self.verbose)
         Base.metadata.create_all(bind=self.engine)
 
-    def insert_test_start_time(self, test_name: str, start_time: int) -> None:
+    def insert_start_time(self, name: str, start_time: int) -> None:
+        """
+        Insert start time only if it doesn't exist.
+
+        Args:
+            name (str): Test/class/module identifier.
+            start_time (int): Start time in seconds since epoch.
+        """
         with Session(bind=self.engine) as db_session:
-            new_table_entry = CnvTestTable(test_name=test_name, start_time=start_time)
-            db_session.add(new_table_entry)
-            db_session.commit()
+            existing_entry = db_session.query(CnvTestTable).filter_by(test_name=name).first()
+
+            if not existing_entry:
+                new_entry = CnvTestTable(test_name=name, start_time=start_time)
+                db_session.add(new_entry)
+                db_session.commit()
+
+    def get_start_time(self, name: str) -> int | None:
+        """
+        Get the start time for a test/class/module.
 
-    def get_test_start_time(self, test_name: str) -> int:
+        Args:
+            name (str): Test/class/module identifier.
+
+        Returns:
+            int | None: Start time in seconds since epoch, or None if not found.
+        """
         with Session(bind=self.engine) as db_session:
-            return (
-                db_session.query(CnvTestTable)
-                .with_entities(CnvTestTable.start_time)
-                .filter_by(test_name=test_name)
-                .one()[0]
+            result = (
+                db_session.query(CnvTestTable).with_entities(CnvTestTable.start_time).filter_by(test_name=name).first()
             )
+            return result[0] if result else None
+
+    def get_start_time_for_collection(self, node: Item | Collector) -> int:
+        """
+        Get test start time based on data_collector_scope marker.
+
+        Determines the appropriate scope (test, class, or module) from the marker,
+        retrieves the start time from the database, and logs the time delta.
+
+        Args:
+            node: Pytest node (Item or Collector).
+
+        Returns:
+            Start time in seconds since epoch, or 0 if not found.
+        """
+        try:
+            # Check data_collector_scope marker
+            scope_marker = node.get_closest_marker(name="data_collector_scope")
+            scope_value = scope_marker.kwargs.get("scope") if scope_marker else None
+
+            name = get_scope_identifier(node=node, scope_value=scope_value)
+            scope_label = scope_value.upper() if scope_value else "TEST"
+
+            test_start_time = self.get_start_time(name=name)
+            if test_start_time:
+                time_delta = int(datetime.datetime.now().strftime("%s")) - test_start_time
+                LOGGER.info(f"[DATA_COLLECTOR] {scope_label} scope: {time_delta}s ({time_delta // 60}m)")
+            else:
+                test_start_time = 0
+                LOGGER.warning(f"[DATA_COLLECTOR] Start time not found for {name}")
+        except Exception as db_exception:
+            test_start_time = 0
+            LOGGER.warning(f"[DATA_COLLECTOR] Error: {db_exception} in accessing database.")
+
+        return test_start_time