e06f895455
Previously the same file name was used for each atomic action, causing overwriting. Changing the logic to write to a different file each time. + Dump JSON in the relevant results directory + Update Elasticsearch indexing status in summary + Add failover to Shaker results also Sample output: https://gist.github.com/smalleni/c25ef05c9b815a8e3299c2fc904908ef Change-Id: Ib74cfd2ecf5c63857548c0ef219e7965b6323d56
102 lines
3.6 KiB
Python
102 lines
3.6 KiB
Python
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
import abc
|
|
import logging
|
|
import os
|
|
import yaml
|
|
|
|
|
|
class WorkloadBase(object):
|
|
__metaclass__ = abc.ABCMeta
|
|
logger = logging.getLogger('browbeat.WorkloadBase')
|
|
success = 0
|
|
failure = 0
|
|
total_tests = 0
|
|
total_scenarios = 0
|
|
browbeat = {}
|
|
|
|
@abc.abstractmethod
|
|
def update_scenarios(self):
|
|
pass
|
|
|
|
@abc.abstractmethod
|
|
def update_tests(self):
|
|
pass
|
|
|
|
@abc.abstractmethod
|
|
def update_pass_tests(self):
|
|
pass
|
|
|
|
@abc.abstractmethod
|
|
def update_fail_tests(self):
|
|
pass
|
|
|
|
def update_total_scenarios(self):
|
|
WorkloadBase.total_scenarios += 1
|
|
|
|
def update_total_tests(self):
|
|
WorkloadBase.total_tests += 1
|
|
|
|
def update_total_pass_tests(self):
|
|
WorkloadBase.success += 1
|
|
|
|
def update_total_fail_tests(self):
|
|
WorkloadBase.failure += 1
|
|
|
|
def workload_logger(self, result_dir, workload):
|
|
base = result_dir.split('/')
|
|
if not os.path.isfile("{}/{}/browbeat-{}-run.log".format(base[0], base[1], workload)):
|
|
file = logging.FileHandler(
|
|
"{}/{}/browbeat-{}-run.log".format(base[0], base[1], workload))
|
|
file.setLevel(logging.DEBUG)
|
|
formatter = logging.Formatter(
|
|
'%(asctime)s - %(name)s - %(levelname)5s - %(message)s')
|
|
file.setFormatter(formatter)
|
|
self.logger.addHandler(file)
|
|
return None
|
|
|
|
def get_time_dict(self, to_time, from_time, benchmark, test_name, workload, status,
|
|
index_status="disabled"):
|
|
time_diff = (to_time - from_time)
|
|
if workload not in WorkloadBase.browbeat:
|
|
WorkloadBase.browbeat[workload] = {}
|
|
if benchmark not in WorkloadBase.browbeat[workload]:
|
|
WorkloadBase.browbeat[workload][benchmark] = {}
|
|
if 'tests' not in WorkloadBase.browbeat[workload][benchmark]:
|
|
WorkloadBase.browbeat[workload][benchmark]['tests'] = []
|
|
if index_status is True:
|
|
index_status = "success"
|
|
elif index_status is False:
|
|
index_status = "failure"
|
|
WorkloadBase.browbeat[workload][benchmark]['tests'].append(
|
|
{'Test name': test_name, 'Time': time_diff, 'Test Status': status,
|
|
'Elasticsearch Indexing': index_status})
|
|
|
|
@staticmethod
|
|
def print_report(result_dir, time_stamp):
|
|
with open(os.path.join(result_dir, time_stamp + '.' + 'report'), 'w') as yaml_file:
|
|
yaml_file.write("Browbeat Report Card\n")
|
|
if not WorkloadBase.browbeat:
|
|
yaml_file.write("No tests were enabled")
|
|
else:
|
|
yaml_file.write(
|
|
yaml.dump(WorkloadBase.browbeat, default_flow_style=False))
|
|
|
|
@staticmethod
|
|
def print_summary():
|
|
WorkloadBase.logger.info("Total scenarios executed:{}".format(
|
|
WorkloadBase.total_scenarios))
|
|
WorkloadBase.logger.info("Total tests executed:{}".format(WorkloadBase.total_tests))
|
|
WorkloadBase.logger.info("Total tests passed:{}".format(WorkloadBase.success))
|
|
WorkloadBase.logger.info("Total tests failed:{}".format(WorkloadBase.failure))
|