1#!/usr/bin/env python3 2# Copyright 2017 gRPC authors. 3# 4# Licensed under the Apache License, Version 2.0 (the "License"); 5# you may not use this file except in compliance with the License. 6# You may obtain a copy of the License at 7# 8# http://www.apache.org/licenses/LICENSE-2.0 9# 10# Unless required by applicable law or agreed to in writing, software 11# distributed under the License is distributed on an "AS IS" BASIS, 12# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 13# See the License for the specific language governing permissions and 14# limitations under the License. 15"""Uploads RBE results to BigQuery""" 16 17import argparse 18import json 19import os 20import ssl 21import sys 22import urllib.error 23import urllib.parse 24import urllib.request 25import uuid 26 27gcp_utils_dir = os.path.abspath( 28 os.path.join(os.path.dirname(__file__), '../../gcp/utils')) 29sys.path.append(gcp_utils_dir) 30import big_query_utils 31 32_DATASET_ID = 'jenkins_test_results' 33_DESCRIPTION = 'Test results from master RBE builds on Kokoro' 34# 365 days in milliseconds 35_EXPIRATION_MS = 365 * 24 * 60 * 60 * 1000 36_PARTITION_TYPE = 'DAY' 37_PROJECT_ID = 'grpc-testing' 38_RESULTS_SCHEMA = [ 39 ('job_name', 'STRING', 'Name of Kokoro job'), 40 ('build_id', 'INTEGER', 'Build ID of Kokoro job'), 41 ('build_url', 'STRING', 'URL of Kokoro build'), 42 ('test_target', 'STRING', 'Bazel target path'), 43 ('test_class_name', 'STRING', 'Name of test class'), 44 ('test_case', 'STRING', 'Name of test case'), 45 ('result', 'STRING', 'Test or build result'), 46 ('timestamp', 'TIMESTAMP', 'Timestamp of test run'), 47 ('duration', 'FLOAT', 'Duration of the test run'), 48] 49_TABLE_ID = 'rbe_test_results' 50 51 52def _get_api_key(): 53 """Returns string with API key to access ResultStore. 54 Intended to be used in Kokoro environment.""" 55 api_key_directory = os.getenv('KOKORO_GFILE_DIR') 56 api_key_file = os.path.join(api_key_directory, 'resultstore_api_key') 57 assert os.path.isfile(api_key_file), 'Must add --api_key arg if not on ' \ 58 'Kokoro or Kokoro environment is not set up properly.' 59 with open(api_key_file, 'r') as f: 60 return f.read().replace('\n', '') 61 62 63def _get_invocation_id(): 64 """Returns String of Bazel invocation ID. Intended to be used in 65 Kokoro environment.""" 66 bazel_id_directory = os.getenv('KOKORO_ARTIFACTS_DIR') 67 bazel_id_file = os.path.join(bazel_id_directory, 'bazel_invocation_ids') 68 assert os.path.isfile(bazel_id_file), 'bazel_invocation_ids file, written ' \ 69 'by RBE initialization script, expected but not found.' 70 with open(bazel_id_file, 'r') as f: 71 return f.read().replace('\n', '') 72 73 74def _parse_test_duration(duration_str): 75 """Parse test duration string in '123.567s' format""" 76 try: 77 if duration_str.endswith('s'): 78 duration_str = duration_str[:-1] 79 return float(duration_str) 80 except: 81 return None 82 83 84def _upload_results_to_bq(rows): 85 """Upload test results to a BQ table. 86 87 Args: 88 rows: A list of dictionaries containing data for each row to insert 89 """ 90 bq = big_query_utils.create_big_query() 91 big_query_utils.create_partitioned_table(bq, 92 _PROJECT_ID, 93 _DATASET_ID, 94 _TABLE_ID, 95 _RESULTS_SCHEMA, 96 _DESCRIPTION, 97 partition_type=_PARTITION_TYPE, 98 expiration_ms=_EXPIRATION_MS) 99 100 max_retries = 3 101 for attempt in range(max_retries): 102 if big_query_utils.insert_rows(bq, _PROJECT_ID, _DATASET_ID, _TABLE_ID, 103 rows): 104 break 105 else: 106 if attempt < max_retries - 1: 107 print('Error uploading result to bigquery, will retry.') 108 else: 109 print( 110 'Error uploading result to bigquery, all attempts failed.') 111 sys.exit(1) 112 113 114def _get_resultstore_data(api_key, invocation_id): 115 """Returns dictionary of test results by querying ResultStore API. 116 Args: 117 api_key: String of ResultStore API key 118 invocation_id: String of ResultStore invocation ID to results from 119 """ 120 all_actions = [] 121 page_token = '' 122 # ResultStore's API returns data on a limited number of tests. When we exceed 123 # that limit, the 'nextPageToken' field is included in the request to get 124 # subsequent data, so keep requesting until 'nextPageToken' field is omitted. 125 while True: 126 req = urllib.request.Request( 127 url= 128 'https://resultstore.googleapis.com/v2/invocations/%s/targets/-/configuredTargets/-/actions?key=%s&pageToken=%s&fields=next_page_token,actions.id,actions.status_attributes,actions.timing,actions.test_action' 129 % (invocation_id, api_key, page_token), 130 headers={'Content-Type': 'application/json'}) 131 ctx_dict = {} 132 if os.getenv("PYTHONHTTPSVERIFY") == "0": 133 ctx = ssl.create_default_context() 134 ctx.check_hostname = False 135 ctx.verify_mode = ssl.CERT_NONE 136 ctx_dict = {"context": ctx} 137 raw_resp = urllib.request.urlopen(req, **ctx_dict).read() 138 decoded_resp = raw_resp if isinstance( 139 raw_resp, str) else raw_resp.decode('utf-8', 'ignore') 140 results = json.loads(decoded_resp) 141 all_actions.extend(results['actions']) 142 if 'nextPageToken' not in results: 143 break 144 page_token = results['nextPageToken'] 145 return all_actions 146 147 148if __name__ == "__main__": 149 # Arguments are necessary if running in a non-Kokoro environment. 150 argp = argparse.ArgumentParser( 151 description= 152 'Fetches results for given RBE invocation and uploads them to BigQuery table.' 153 ) 154 argp.add_argument('--api_key', 155 default='', 156 type=str, 157 help='The API key to read from ResultStore API') 158 argp.add_argument('--invocation_id', 159 default='', 160 type=str, 161 help='UUID of bazel invocation to fetch.') 162 argp.add_argument('--bq_dump_file', 163 default=None, 164 type=str, 165 help='Dump JSON data to file just before uploading') 166 argp.add_argument('--resultstore_dump_file', 167 default=None, 168 type=str, 169 help='Dump JSON data as received from ResultStore API') 170 argp.add_argument('--skip_upload', 171 default=False, 172 action='store_const', 173 const=True, 174 help='Skip uploading to bigquery') 175 args = argp.parse_args() 176 177 api_key = args.api_key or _get_api_key() 178 invocation_id = args.invocation_id or _get_invocation_id() 179 resultstore_actions = _get_resultstore_data(api_key, invocation_id) 180 181 if args.resultstore_dump_file: 182 with open(args.resultstore_dump_file, 'w') as f: 183 json.dump(resultstore_actions, f, indent=4, sort_keys=True) 184 print( 185 ('Dumped resultstore data to file %s' % args.resultstore_dump_file)) 186 187 # google.devtools.resultstore.v2.Action schema: 188 # https://github.com/googleapis/googleapis/blob/master/google/devtools/resultstore/v2/action.proto 189 bq_rows = [] 190 for index, action in enumerate(resultstore_actions): 191 # Filter out non-test related data, such as build results. 192 if 'testAction' not in action: 193 continue 194 # Some test results contain the fileProcessingErrors field, which indicates 195 # an issue with parsing results individual test cases. 196 if 'fileProcessingErrors' in action: 197 test_cases = [{ 198 'testCase': { 199 'caseName': str(action['id']['actionId']), 200 } 201 }] 202 # Test timeouts have a different dictionary structure compared to pass and 203 # fail results. 204 elif action['statusAttributes']['status'] == 'TIMED_OUT': 205 test_cases = [{ 206 'testCase': { 207 'caseName': str(action['id']['actionId']), 208 'timedOut': True 209 } 210 }] 211 # When RBE believes its infrastructure is failing, it will abort and 212 # mark running tests as UNKNOWN. These infrastructure failures may be 213 # related to our tests, so we should investigate if specific tests are 214 # repeatedly being marked as UNKNOWN. 215 elif action['statusAttributes']['status'] == 'UNKNOWN': 216 test_cases = [{ 217 'testCase': { 218 'caseName': str(action['id']['actionId']), 219 'unknown': True 220 } 221 }] 222 # Take the timestamp from the previous action, which should be 223 # a close approximation. 224 action['timing'] = { 225 'startTime': 226 resultstore_actions[index - 1]['timing']['startTime'] 227 } 228 elif 'testSuite' not in action['testAction']: 229 continue 230 elif 'tests' not in action['testAction']['testSuite']: 231 continue 232 else: 233 test_cases = [] 234 for tests_item in action['testAction']['testSuite']['tests']: 235 test_cases += tests_item['testSuite']['tests'] 236 for test_case in test_cases: 237 if any(s in test_case['testCase'] for s in ['errors', 'failures']): 238 result = 'FAILED' 239 elif 'timedOut' in test_case['testCase']: 240 result = 'TIMEOUT' 241 elif 'unknown' in test_case['testCase']: 242 result = 'UNKNOWN' 243 else: 244 result = 'PASSED' 245 try: 246 bq_rows.append({ 247 'insertId': str(uuid.uuid4()), 248 'json': { 249 'job_name': 250 os.getenv('KOKORO_JOB_NAME'), 251 'build_id': 252 os.getenv('KOKORO_BUILD_NUMBER'), 253 'build_url': 254 'https://source.cloud.google.com/results/invocations/%s' 255 % invocation_id, 256 'test_target': 257 action['id']['targetId'], 258 'test_class_name': 259 test_case['testCase'].get('className', ''), 260 'test_case': 261 test_case['testCase']['caseName'], 262 'result': 263 result, 264 'timestamp': 265 action['timing']['startTime'], 266 'duration': 267 _parse_test_duration(action['timing']['duration']), 268 } 269 }) 270 except Exception as e: 271 print(('Failed to parse test result. Error: %s' % str(e))) 272 print((json.dumps(test_case, indent=4))) 273 bq_rows.append({ 274 'insertId': str(uuid.uuid4()), 275 'json': { 276 'job_name': 277 os.getenv('KOKORO_JOB_NAME'), 278 'build_id': 279 os.getenv('KOKORO_BUILD_NUMBER'), 280 'build_url': 281 'https://source.cloud.google.com/results/invocations/%s' 282 % invocation_id, 283 'test_target': 284 action['id']['targetId'], 285 'test_class_name': 286 'N/A', 287 'test_case': 288 'N/A', 289 'result': 290 'UNPARSEABLE', 291 'timestamp': 292 'N/A', 293 } 294 }) 295 296 if args.bq_dump_file: 297 with open(args.bq_dump_file, 'w') as f: 298 json.dump(bq_rows, f, indent=4, sort_keys=True) 299 print(('Dumped BQ data to file %s' % args.bq_dump_file)) 300 301 if not args.skip_upload: 302 # BigQuery sometimes fails with large uploads, so batch 1,000 rows at a time. 303 MAX_ROWS = 1000 304 for i in range(0, len(bq_rows), MAX_ROWS): 305 _upload_results_to_bq(bq_rows[i:i + MAX_ROWS]) 306 else: 307 print('Skipped upload to bigquery.') 308