import random import re from datetime import datetime, timezone import allure import pytest from frostfs_testlib import reporter from frostfs_testlib.steps.metrics import get_metrics_value from frostfs_testlib.storage.cluster import ClusterNode from frostfs_testlib.storage.controllers.cluster_state_controller import ClusterStateController from frostfs_testlib.storage.controllers.state_managers.config_state_manager import ConfigStateManager from frostfs_testlib.storage.dataclasses.frostfs_services import StorageNode from frostfs_testlib.testing.cluster_test_base import ClusterTestBase from frostfs_testlib.testing.test_control import wait_for_success @pytest.mark.nightly @pytest.mark.metrics class TestLogsMetrics(ClusterTestBase): @pytest.fixture def revert_all(self, cluster_state_controller: ClusterStateController): yield cluster_state_controller.manager(ConfigStateManager).revert_all() def restart_storage_service(self, cluster_state_controller: ClusterStateController) -> datetime: config_manager = cluster_state_controller.manager(ConfigStateManager) config_manager.csc.stop_services_of_type(StorageNode) restart_time = datetime.now(timezone.utc) config_manager.csc.start_services_of_type(StorageNode) return restart_time @wait_for_success(interval=10) def check_metrics_in_node(self, cluster_node: ClusterNode, restart_time: datetime, log_priority: str = None, **metrics_greps): current_time = datetime.now(timezone.utc) counter_metrics = get_metrics_value(cluster_node, **metrics_greps) counter_logs = self.get_count_logs_by_level(cluster_node, metrics_greps.get("level"), restart_time, current_time, log_priority) assert counter_logs == pytest.approx( counter_metrics, rel=0.02 ), f"counter_logs: {counter_logs}, counter_metrics: {counter_metrics} in node: {cluster_node}" @staticmethod def get_count_logs_by_level(cluster_node: ClusterNode, log_level: str, after_time: datetime, until_time: datetime, log_priority: str): count_logs = 0 try: logs = cluster_node.host.get_filtered_logs( log_level, unit="frostfs-storage", since=after_time, until=until_time, priority=log_priority ) result = re.findall(rf":\s+{log_level}\s+", logs) count_logs += len(result) except RuntimeError as e: ... return count_logs @allure.title("Metrics for the log counter") def test_log_counter_metrics(self, cluster_state_controller: ClusterStateController, revert_all): restart_time = self.restart_storage_service(cluster_state_controller) with reporter.step("Select random node"): node = random.choice(self.cluster.cluster_nodes) with reporter.step(f"Check metrics count logs with level 'info'"): self.check_metrics_in_node( node, restart_time, log_priority="6..6", command="frostfs_node_logger_entry_count", level="info", dropped="false", ) with reporter.step(f"Check metrics count logs with level 'error'"): self.check_metrics_in_node(node, restart_time, command="frostfs_node_logger_entry_count", level="error", dropped="false")