mirror of https://github.com/apache/kafka.git
952 lines
39 KiB
Python
952 lines
39 KiB
Python
# Licensed to the Apache Software Foundation (ASF) under one or more
|
|
# contributor license agreements. See the NOTICE file distributed with
|
|
# this work for additional information regarding copyright ownership.
|
|
# The ASF licenses this file to You 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 os
|
|
import requests
|
|
import json
|
|
from dataclasses import dataclass, field, asdict
|
|
from typing import Dict, List, Tuple, Optional
|
|
from datetime import datetime, timedelta
|
|
import pytz # Add this import for timezone handling
|
|
from collections import defaultdict
|
|
import time
|
|
import logging
|
|
import concurrent.futures # Add this import at the top
|
|
import pathlib
|
|
import pickle
|
|
from abc import ABC, abstractmethod
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
@dataclass
|
|
class TestOutcome:
|
|
passed: int
|
|
failed: int
|
|
skipped: int
|
|
flaky: int
|
|
not_selected: int = field(metadata={'name': 'notSelected'})
|
|
total: int
|
|
|
|
@dataclass
|
|
class BuildInfo:
|
|
id: str
|
|
timestamp: datetime
|
|
duration: int
|
|
has_failed: bool
|
|
|
|
@dataclass
|
|
class TestTimelineEntry:
|
|
build_id: str
|
|
timestamp: datetime
|
|
outcome: str # "passed", "failed", "flaky", etc.
|
|
|
|
@dataclass
|
|
class TestResult:
|
|
name: str
|
|
outcome_distribution: TestOutcome
|
|
first_seen: datetime
|
|
timeline: List[TestTimelineEntry] = field(default_factory=list)
|
|
recent_failure_rate: float = 0.0 # Added to track recent failure trends
|
|
|
|
@dataclass
|
|
class TestContainerResult:
|
|
build_id: str
|
|
outcome: str
|
|
timestamp: Optional[datetime] = None
|
|
|
|
@dataclass
|
|
class TestCaseResult(TestResult):
|
|
"""Extends TestResult to include container-specific information"""
|
|
container_name: str = ""
|
|
|
|
@dataclass
|
|
class BuildCache:
|
|
last_update: datetime
|
|
builds: Dict[str, 'BuildInfo']
|
|
|
|
def to_dict(self):
|
|
return {
|
|
'last_update': self.last_update.isoformat(),
|
|
'builds': {k: asdict(v) for k, v in self.builds.items()}
|
|
}
|
|
|
|
@classmethod
|
|
def from_dict(cls, data: dict) -> 'BuildCache':
|
|
return cls(
|
|
last_update=datetime.fromisoformat(data['last_update']),
|
|
builds={k: BuildInfo(**v) for k, v in data['builds'].items()}
|
|
)
|
|
|
|
class CacheProvider(ABC):
|
|
@abstractmethod
|
|
def get_cache(self) -> Optional[BuildCache]:
|
|
pass
|
|
|
|
@abstractmethod
|
|
def save_cache(self, cache: BuildCache):
|
|
pass
|
|
|
|
class LocalCacheProvider(CacheProvider):
|
|
def __init__(self, cache_dir: str = None):
|
|
if cache_dir is None:
|
|
cache_dir = os.path.join(os.path.expanduser("~"), ".develocity_cache")
|
|
self.cache_file = os.path.join(cache_dir, "build_cache.pkl")
|
|
os.makedirs(cache_dir, exist_ok=True)
|
|
|
|
def get_cache(self) -> Optional[BuildCache]:
|
|
try:
|
|
if os.path.exists(self.cache_file):
|
|
with open(self.cache_file, 'rb') as f:
|
|
return pickle.load(f)
|
|
except Exception as e:
|
|
logger.warning(f"Failed to load local cache: {e}")
|
|
return None
|
|
|
|
def save_cache(self, cache: BuildCache):
|
|
try:
|
|
with open(self.cache_file, 'wb') as f:
|
|
pickle.dump(cache, f)
|
|
except Exception as e:
|
|
logger.warning(f"Failed to save local cache: {e}")
|
|
|
|
class GitHubActionsCacheProvider(CacheProvider):
|
|
def __init__(self):
|
|
self.cache_key = "develocity-build-cache"
|
|
|
|
def get_cache(self) -> Optional[BuildCache]:
|
|
try:
|
|
# Check if running in GitHub Actions
|
|
if not os.environ.get('GITHUB_ACTIONS'):
|
|
return None
|
|
|
|
cache_path = os.environ.get('GITHUB_WORKSPACE', '')
|
|
cache_file = os.path.join(cache_path, self.cache_key + '.json')
|
|
|
|
if os.path.exists(cache_file):
|
|
with open(cache_file, 'r') as f:
|
|
data = json.load(f)
|
|
return BuildCache.from_dict(data)
|
|
except Exception as e:
|
|
logger.warning(f"Failed to load GitHub Actions cache: {e}")
|
|
return None
|
|
|
|
def save_cache(self, cache: BuildCache):
|
|
try:
|
|
if not os.environ.get('GITHUB_ACTIONS'):
|
|
return
|
|
|
|
cache_path = os.environ.get('GITHUB_WORKSPACE', '')
|
|
cache_file = os.path.join(cache_path, self.cache_key + '.json')
|
|
|
|
with open(cache_file, 'w') as f:
|
|
json.dump(cache.to_dict(), f)
|
|
except Exception as e:
|
|
logger.warning(f"Failed to save GitHub Actions cache: {e}")
|
|
|
|
class TestAnalyzer:
|
|
def __init__(self, base_url: str, auth_token: str):
|
|
self.base_url = base_url
|
|
self.headers = {
|
|
'Authorization': f'Bearer {auth_token}',
|
|
'Accept': 'application/json'
|
|
}
|
|
self.default_chunk_size = timedelta(days=14)
|
|
self.api_retry_delay = 2 # seconds
|
|
self.max_api_retries = 3
|
|
|
|
# Initialize cache providers
|
|
self.cache_providers = [
|
|
GitHubActionsCacheProvider(),
|
|
LocalCacheProvider()
|
|
]
|
|
self.build_cache = None
|
|
self._load_cache()
|
|
|
|
def _load_cache(self):
|
|
"""Load cache from the first available provider"""
|
|
for provider in self.cache_providers:
|
|
cache = provider.get_cache()
|
|
if cache is not None:
|
|
self.build_cache = cache
|
|
logger.info(f"Loaded cache from {provider.__class__.__name__}")
|
|
return
|
|
logger.info("No existing cache found")
|
|
|
|
def _save_cache(self):
|
|
"""Save cache to all providers"""
|
|
if self.build_cache:
|
|
for provider in self.cache_providers:
|
|
provider.save_cache(self.build_cache)
|
|
logger.info(f"Saved cache to {provider.__class__.__name__}")
|
|
|
|
def build_query(self, project: str, chunk_start: datetime, chunk_end: datetime, test_type: str) -> str:
|
|
"""
|
|
Constructs the query string to be used in both build info and test containers API calls.
|
|
|
|
Args:
|
|
project: The project name.
|
|
chunk_start: The start datetime for the chunk.
|
|
chunk_end: The end datetime for the chunk.
|
|
test_type: The type of tests to query.
|
|
|
|
Returns:
|
|
A formatted query string.
|
|
"""
|
|
return f'project:{project} buildStartTime:[{chunk_start.isoformat()} TO {chunk_end.isoformat()}] gradle.requestedTasks:{test_type}'
|
|
|
|
def process_chunk(
|
|
self,
|
|
chunk_start: datetime,
|
|
chunk_end: datetime,
|
|
project: str,
|
|
test_type: str,
|
|
remaining_build_ids: set,
|
|
max_builds_per_request: int
|
|
) -> Dict[str, BuildInfo]:
|
|
"""Helper method to process a single chunk of build information"""
|
|
chunk_builds = {}
|
|
|
|
# Use the helper method to build the query
|
|
query = self.build_query(project, chunk_start, chunk_end, test_type)
|
|
|
|
# Initialize pagination for this chunk
|
|
from_build = None
|
|
continue_chunk = True
|
|
|
|
while continue_chunk and remaining_build_ids:
|
|
query_params = {
|
|
'query': query,
|
|
'models': ['gradle-attributes'],
|
|
'allModels': 'false',
|
|
'maxBuilds': max_builds_per_request,
|
|
'reverse': 'false',
|
|
'fromInstant': int(chunk_start.timestamp() * 1000)
|
|
}
|
|
|
|
if from_build:
|
|
query_params['fromBuild'] = from_build
|
|
|
|
for attempt in range(self.max_api_retries):
|
|
try:
|
|
response = requests.get(
|
|
f'{self.base_url}/api/builds',
|
|
headers=self.headers,
|
|
params=query_params,
|
|
timeout=(5, 30)
|
|
)
|
|
response.raise_for_status()
|
|
break
|
|
except requests.exceptions.Timeout:
|
|
if attempt == self.max_api_retries - 1:
|
|
raise
|
|
time.sleep(self.api_retry_delay * (attempt + 1))
|
|
except requests.exceptions.RequestException:
|
|
raise
|
|
|
|
response_json = response.json()
|
|
|
|
if not response_json:
|
|
break
|
|
|
|
for build in response_json:
|
|
build_id = build['id']
|
|
|
|
if 'models' in build and 'gradleAttributes' in build['models']:
|
|
gradle_attrs = build['models']['gradleAttributes']
|
|
if 'model' in gradle_attrs:
|
|
attrs = gradle_attrs['model']
|
|
build_timestamp = datetime.fromtimestamp(attrs['buildStartTime'] / 1000, pytz.UTC)
|
|
|
|
if build_timestamp >= chunk_end:
|
|
continue_chunk = False
|
|
break
|
|
|
|
if build_id in remaining_build_ids:
|
|
if 'problem' not in gradle_attrs:
|
|
chunk_builds[build_id] = BuildInfo(
|
|
id=build_id,
|
|
timestamp=build_timestamp,
|
|
duration=attrs.get('buildDuration'),
|
|
has_failed=attrs.get('hasFailed', False)
|
|
)
|
|
|
|
if continue_chunk and response_json:
|
|
from_build = response_json[-1]['id']
|
|
else:
|
|
continue_chunk = False
|
|
|
|
time.sleep(0.5) # Rate limiting between pagination requests
|
|
|
|
return chunk_builds
|
|
|
|
def get_build_info(self, build_ids: List[str], project: str, test_type: str, query_days: int) -> Dict[str, BuildInfo]:
|
|
builds = {}
|
|
max_builds_per_request = 100
|
|
cutoff_date = datetime.now(pytz.UTC) - timedelta(days=query_days)
|
|
|
|
# Get builds from cache if available
|
|
if self.build_cache:
|
|
cached_builds = self.build_cache.builds
|
|
cached_cutoff = self.build_cache.last_update - timedelta(days=query_days)
|
|
|
|
# Use cached data for builds within the cache period
|
|
for build_id in build_ids:
|
|
if build_id in cached_builds:
|
|
build = cached_builds[build_id]
|
|
if build.timestamp >= cached_cutoff:
|
|
builds[build_id] = build
|
|
|
|
# Update cutoff date to only fetch new data
|
|
cutoff_date = self.build_cache.last_update
|
|
logger.info(f"Using cached data up to {cutoff_date.isoformat()}")
|
|
|
|
# Remove already found builds from the search list
|
|
build_ids = [bid for bid in build_ids if bid not in builds]
|
|
|
|
if not build_ids:
|
|
logger.info("All builds found in cache")
|
|
return builds
|
|
|
|
# Fetch remaining builds from API
|
|
remaining_build_ids = set(build_ids)
|
|
current_time = datetime.now(pytz.UTC)
|
|
chunk_size = self.default_chunk_size
|
|
|
|
# Create time chunks
|
|
chunks = []
|
|
chunk_start = cutoff_date
|
|
while chunk_start < current_time:
|
|
chunk_end = min(chunk_start + chunk_size, current_time)
|
|
chunks.append((chunk_start, chunk_end))
|
|
chunk_start = chunk_end
|
|
|
|
total_start_time = time.time()
|
|
|
|
# Process chunks in parallel
|
|
with concurrent.futures.ThreadPoolExecutor(max_workers=10) as executor:
|
|
future_to_chunk = {
|
|
executor.submit(
|
|
self.process_chunk,
|
|
chunk[0],
|
|
chunk[1],
|
|
project,
|
|
test_type,
|
|
remaining_build_ids.copy(),
|
|
max_builds_per_request
|
|
): chunk for chunk in chunks
|
|
}
|
|
|
|
for future in concurrent.futures.as_completed(future_to_chunk):
|
|
try:
|
|
chunk_builds = future.result()
|
|
builds.update(chunk_builds)
|
|
remaining_build_ids -= set(chunk_builds.keys())
|
|
except Exception as e:
|
|
logger.error(f"Chunk processing generated an exception: {str(e)}")
|
|
|
|
total_duration = time.time() - total_start_time
|
|
logger.info(
|
|
f"\nBuild Info Performance:"
|
|
f"\n Total Duration: {total_duration:.2f}s"
|
|
f"\n Builds Retrieved: {len(builds)}"
|
|
f"\n Builds Not Found: {len(remaining_build_ids)}"
|
|
)
|
|
|
|
# Update cache with new data
|
|
if builds:
|
|
if not self.build_cache:
|
|
self.build_cache = BuildCache(current_time, {})
|
|
self.build_cache.builds.update(builds)
|
|
self.build_cache.last_update = current_time
|
|
self._save_cache()
|
|
|
|
return builds
|
|
|
|
def get_test_results(self, project: str, threshold_days: int, test_type: str = "quarantinedTest",
|
|
outcomes: List[str] = None) -> List[TestResult]:
|
|
"""Fetch test results with timeline information"""
|
|
if outcomes is None:
|
|
outcomes = ["failed", "flaky"]
|
|
|
|
logger.debug(f"Fetching test results for project {project}, last {threshold_days} days")
|
|
|
|
end_time = datetime.now(pytz.UTC)
|
|
start_time = end_time - timedelta(days=threshold_days)
|
|
|
|
all_results = {}
|
|
build_ids = set()
|
|
test_container_results = defaultdict(list)
|
|
|
|
chunk_size = self.default_chunk_size
|
|
chunk_start = start_time
|
|
|
|
while chunk_start < end_time:
|
|
chunk_end = min(chunk_start + chunk_size, end_time)
|
|
logger.debug(f"Processing chunk: {chunk_start} to {chunk_end}")
|
|
|
|
# Use the helper method to build the query
|
|
query = self.build_query(project, chunk_start, chunk_end, test_type)
|
|
|
|
query_params = {
|
|
'query': query,
|
|
'testOutcomes': outcomes,
|
|
'container': '*',
|
|
'include': ['buildScanIds'] # Explicitly request build scan IDs
|
|
}
|
|
|
|
response = requests.get(
|
|
f'{self.base_url}/api/tests/containers',
|
|
headers=self.headers,
|
|
params=query_params
|
|
)
|
|
response.raise_for_status()
|
|
|
|
for test in response.json()['content']:
|
|
test_name = test['name']
|
|
logger.debug(f"Processing test: {test_name}")
|
|
|
|
if test_name not in all_results:
|
|
outcome_data = test['outcomeDistribution']
|
|
if 'notSelected' in outcome_data:
|
|
outcome_data['not_selected'] = outcome_data.pop('notSelected')
|
|
outcome = TestOutcome(**outcome_data)
|
|
all_results[test_name] = TestResult(test_name, outcome, chunk_start)
|
|
|
|
# Collect build IDs by outcome
|
|
if 'buildScanIdsByOutcome' in test:
|
|
scan_ids = test['buildScanIdsByOutcome']
|
|
|
|
for outcome, ids in scan_ids.items():
|
|
if ids: # Only process if we have IDs
|
|
for build_id in ids:
|
|
build_ids.add(build_id)
|
|
test_container_results[test_name].append(
|
|
TestContainerResult(build_id=build_id, outcome=outcome)
|
|
)
|
|
|
|
chunk_start = chunk_end
|
|
|
|
logger.debug(f"Total unique build IDs collected: {len(build_ids)}")
|
|
|
|
# Fetch build information using the updated get_build_info method
|
|
builds = self.get_build_info(list(build_ids), project, test_type, threshold_days)
|
|
logger.debug(f"Retrieved {len(builds)} builds from API")
|
|
logger.debug(f"Retrieved build IDs: {sorted(builds.keys())}")
|
|
|
|
# Update test results with timeline information
|
|
for test_name, result in all_results.items():
|
|
logger.debug(f"\nProcessing timeline for test: {test_name}")
|
|
timeline = []
|
|
for container_result in test_container_results[test_name]:
|
|
logger.debug(f"Processing container result: {container_result}")
|
|
if container_result.build_id in builds:
|
|
build_info = builds[container_result.build_id]
|
|
timeline.append(TestTimelineEntry(
|
|
build_id=container_result.build_id,
|
|
timestamp=build_info.timestamp,
|
|
outcome=container_result.outcome
|
|
))
|
|
else:
|
|
logger.warning(f"Build ID {container_result.build_id} not found in builds response")
|
|
|
|
# Sort timeline by timestamp
|
|
result.timeline = sorted(timeline, key=lambda x: x.timestamp)
|
|
logger.debug(f"Final timeline entries for {test_name}: {len(result.timeline)}")
|
|
|
|
# Calculate recent failure rate
|
|
recent_cutoff = datetime.now(pytz.UTC) - timedelta(days=30)
|
|
recent_runs = [t for t in timeline if t.timestamp >= recent_cutoff]
|
|
if recent_runs:
|
|
recent_failures = sum(1 for t in recent_runs if t.outcome in ('failed', 'flaky'))
|
|
result.recent_failure_rate = recent_failures / len(recent_runs)
|
|
|
|
return list(all_results.values())
|
|
|
|
def get_defective_tests(self, results: List[TestResult]) -> Dict[str, TestResult]:
|
|
"""
|
|
Analyze test results to find defective tests (failed or flaky)
|
|
"""
|
|
defective_tests = {}
|
|
|
|
for result in results:
|
|
if result.outcome_distribution.failed > 0 or result.outcome_distribution.flaky > 0:
|
|
defective_tests[result.name] = result
|
|
|
|
return defective_tests
|
|
|
|
def get_long_quarantined_tests(self, results: List[TestResult], quarantine_threshold_days: int = 60) -> Dict[str, TestResult]:
|
|
"""
|
|
Find tests that have been quarantined longer than the threshold.
|
|
These are candidates for removal or rewriting.
|
|
|
|
Args:
|
|
results: List of test results
|
|
quarantine_threshold_days: Number of days after which a quarantined test should be considered for removal/rewrite
|
|
"""
|
|
long_quarantined = {}
|
|
current_time = datetime.now(pytz.UTC)
|
|
|
|
for result in results:
|
|
days_quarantined = (current_time - result.first_seen).days
|
|
if days_quarantined >= quarantine_threshold_days:
|
|
long_quarantined[result.name] = (result, days_quarantined)
|
|
|
|
return long_quarantined
|
|
|
|
def get_problematic_quarantined_tests(
|
|
self,
|
|
results: List[TestResult],
|
|
quarantine_threshold_days: int = 60,
|
|
min_failure_rate: float = 0.3,
|
|
recent_failure_threshold: float = 0.5
|
|
) -> Dict[str, Dict]:
|
|
"""Enhanced version that includes test case details"""
|
|
problematic_tests = {}
|
|
current_time = datetime.now(pytz.UTC)
|
|
chunk_start = current_time - timedelta(days=7) # Last 7 days for test cases
|
|
|
|
for result in results:
|
|
days_quarantined = (current_time - result.first_seen).days
|
|
if days_quarantined >= quarantine_threshold_days:
|
|
total_runs = result.outcome_distribution.total
|
|
if total_runs > 0:
|
|
problem_runs = result.outcome_distribution.failed + result.outcome_distribution.flaky
|
|
failure_rate = problem_runs / total_runs
|
|
|
|
if failure_rate >= min_failure_rate or result.recent_failure_rate >= recent_failure_threshold:
|
|
# Get detailed test case information
|
|
try:
|
|
test_cases = self.get_test_case_details(
|
|
result.name,
|
|
"kafka",
|
|
chunk_start,
|
|
current_time,
|
|
test_type="quarantinedTest"
|
|
)
|
|
|
|
problematic_tests[result.name] = {
|
|
'container_result': result,
|
|
'days_quarantined': days_quarantined,
|
|
'failure_rate': failure_rate,
|
|
'recent_failure_rate': result.recent_failure_rate,
|
|
'test_cases': test_cases
|
|
}
|
|
except Exception as e:
|
|
logger.error(f"Error getting test case details for {result.name}: {str(e)}")
|
|
|
|
return problematic_tests
|
|
|
|
def get_test_case_details(
|
|
self,
|
|
container_name: str,
|
|
project: str,
|
|
chunk_start: datetime,
|
|
chunk_end: datetime,
|
|
test_type: str = "quarantinedTest"
|
|
) -> List[TestCaseResult]:
|
|
"""
|
|
Fetch detailed test case results for a specific container.
|
|
|
|
Args:
|
|
container_name: Name of the test container
|
|
project: The project name
|
|
chunk_start: Start time for the query
|
|
chunk_end: End time for the query
|
|
test_type: Type of tests to query (default: "quarantinedTest")
|
|
"""
|
|
# Use the helper method to build the query, similar to get_test_results
|
|
query = self.build_query(project, chunk_start, chunk_end, test_type)
|
|
|
|
query_params = {
|
|
'query': query,
|
|
'testOutcomes': ['failed', 'flaky'],
|
|
'container': container_name,
|
|
'include': ['buildScanIds'], # Explicitly request build scan IDs
|
|
'limit': 1000
|
|
}
|
|
|
|
try:
|
|
response = requests.get(
|
|
f'{self.base_url}/api/tests/cases',
|
|
headers=self.headers,
|
|
params=query_params
|
|
)
|
|
response.raise_for_status()
|
|
|
|
test_cases = []
|
|
content = response.json().get('content', [])
|
|
|
|
# Collect all build IDs first
|
|
build_ids = set()
|
|
for test in content:
|
|
if 'buildScanIdsByOutcome' in test:
|
|
for outcome_type, ids in test['buildScanIdsByOutcome'].items():
|
|
build_ids.update(ids)
|
|
|
|
# Get build info for all build IDs
|
|
builds = self.get_build_info(list(build_ids), project, test_type, 7) # 7 days for test cases
|
|
|
|
for test in content:
|
|
outcome_data = test['outcomeDistribution']
|
|
if 'notSelected' in outcome_data:
|
|
outcome_data['not_selected'] = outcome_data.pop('notSelected')
|
|
outcome = TestOutcome(**outcome_data)
|
|
|
|
test_case = TestCaseResult(
|
|
name=test['name'],
|
|
outcome_distribution=outcome,
|
|
first_seen=chunk_start,
|
|
container_name=container_name
|
|
)
|
|
|
|
# Add build information with proper timestamps
|
|
if 'buildScanIdsByOutcome' in test:
|
|
for outcome_type, build_ids in test['buildScanIdsByOutcome'].items():
|
|
for build_id in build_ids:
|
|
if build_id in builds:
|
|
build_info = builds[build_id]
|
|
test_case.timeline.append(
|
|
TestTimelineEntry(
|
|
build_id=build_id,
|
|
timestamp=build_info.timestamp,
|
|
outcome=outcome_type
|
|
)
|
|
)
|
|
else:
|
|
logger.warning(f"Build ID {build_id} not found for test case {test['name']}")
|
|
|
|
# Sort timeline by timestamp
|
|
test_case.timeline.sort(key=lambda x: x.timestamp)
|
|
test_cases.append(test_case)
|
|
|
|
return test_cases
|
|
|
|
except requests.exceptions.RequestException as e:
|
|
logger.error(f"Error fetching test case details for {container_name}: {str(e)}")
|
|
raise
|
|
|
|
def get_flaky_test_regressions(self, project: str, results: List[TestResult],
|
|
recent_days: int = 7, min_flaky_rate: float = 0.2) -> Dict[str, Dict]:
|
|
"""
|
|
Identify tests that have recently started showing flaky behavior.
|
|
|
|
Args:
|
|
project: The project name
|
|
results: List of test results
|
|
recent_days: Number of days to consider for recent behavior
|
|
min_flaky_rate: Minimum flaky rate to consider a test as problematic
|
|
"""
|
|
flaky_regressions = {}
|
|
current_time = datetime.now(pytz.UTC)
|
|
recent_cutoff = current_time - timedelta(days=recent_days)
|
|
|
|
for result in results:
|
|
# Skip tests with no timeline data
|
|
if not result.timeline:
|
|
continue
|
|
|
|
# Split timeline into recent and historical periods
|
|
recent_entries = [t for t in result.timeline if t.timestamp >= recent_cutoff]
|
|
historical_entries = [t for t in result.timeline if t.timestamp < recent_cutoff]
|
|
|
|
if not recent_entries or not historical_entries:
|
|
continue
|
|
|
|
# Calculate flaky rates
|
|
recent_flaky = sum(1 for t in recent_entries if t.outcome == 'flaky')
|
|
recent_total = len(recent_entries)
|
|
recent_flaky_rate = recent_flaky / recent_total if recent_total > 0 else 0
|
|
|
|
historical_flaky = sum(1 for t in historical_entries if t.outcome == 'flaky')
|
|
historical_total = len(historical_entries)
|
|
historical_flaky_rate = historical_flaky / historical_total if historical_total > 0 else 0
|
|
|
|
# Check if there's a significant increase in flakiness
|
|
if recent_flaky_rate >= min_flaky_rate and recent_flaky_rate > historical_flaky_rate * 1.5:
|
|
flaky_regressions[result.name] = {
|
|
'result': result,
|
|
'recent_flaky_rate': recent_flaky_rate,
|
|
'historical_flaky_rate': historical_flaky_rate,
|
|
'recent_executions': recent_entries,
|
|
'historical_executions': historical_entries
|
|
}
|
|
|
|
return flaky_regressions
|
|
|
|
def get_cleared_tests(self, project: str, results: List[TestResult],
|
|
success_threshold: float = 0.7, min_executions: int = 5) -> Dict[str, Dict]:
|
|
"""
|
|
Identify quarantined tests that are consistently passing and could be cleared.
|
|
|
|
Args:
|
|
project: The project name
|
|
results: List of test results
|
|
success_threshold: Required percentage of successful builds to be considered cleared
|
|
min_executions: Minimum number of executions required to make a determination
|
|
"""
|
|
cleared_tests = {}
|
|
current_time = datetime.now(pytz.UTC)
|
|
|
|
for result in results:
|
|
# Only consider tests with sufficient recent executions
|
|
recent_executions = result.timeline
|
|
if len(recent_executions) < min_executions:
|
|
continue
|
|
|
|
# Calculate success rate
|
|
successful_runs = sum(1 for t in recent_executions
|
|
if t.outcome == 'passed')
|
|
success_rate = successful_runs / len(recent_executions)
|
|
|
|
# Check if the test meets clearing criteria
|
|
if success_rate >= success_threshold:
|
|
# Verify no recent failures or flaky behavior
|
|
has_recent_issues = any(t.outcome in ['failed', 'flaky']
|
|
for t in recent_executions[-min_executions:])
|
|
|
|
if not has_recent_issues:
|
|
cleared_tests[result.name] = {
|
|
'result': result,
|
|
'success_rate': success_rate,
|
|
'total_executions': len(recent_executions),
|
|
'successful_runs': successful_runs,
|
|
'recent_executions': recent_executions[-min_executions:]
|
|
}
|
|
|
|
return cleared_tests
|
|
|
|
def print_summary(problematic_tests: Dict[str, Dict], flaky_regressions: Dict[str, Dict]):
|
|
"""Print a summary of the most problematic tests at the top of the report"""
|
|
print("\n## Summary of Most Problematic Tests")
|
|
|
|
# Combine and sort all test cases by failure rate
|
|
all_problem_cases = []
|
|
|
|
# Process problematic quarantined tests
|
|
if len(problematic_tests) > 0:
|
|
print(f"Found {len(problematic_tests)} tests that have been quarantined for a while and are still flaky.")
|
|
for full_class_name, details in problematic_tests.items():
|
|
for test_case in details['test_cases']:
|
|
total_runs = test_case.outcome_distribution.total
|
|
method_name = test_case.name.split('.')[-1]
|
|
if total_runs > 0:
|
|
failure_rate = (test_case.outcome_distribution.failed +
|
|
test_case.outcome_distribution.flaky) / total_runs
|
|
all_problem_cases.append({
|
|
'class': full_class_name,
|
|
'method': method_name,
|
|
'failure_rate': failure_rate,
|
|
'total_runs': total_runs
|
|
})
|
|
|
|
# Process flaky regressions
|
|
if len(flaky_regressions) > 0:
|
|
print(f"Found {len(flaky_regressions)} tests that have started recently failing.")
|
|
for test_name, details in flaky_regressions.items():
|
|
all_problem_cases.append({
|
|
'class': test_name,
|
|
'method': 'N/A', # Flaky regressions are at class level
|
|
'failure_rate': details['recent_flaky_rate'],
|
|
'total_runs': len(details['recent_executions'])
|
|
})
|
|
|
|
# Sort by failure rate descending
|
|
sorted_cases = sorted(all_problem_cases,
|
|
key=lambda x: x['failure_rate'],
|
|
reverse=True)
|
|
|
|
# Group by class
|
|
by_class = {}
|
|
for case in sorted_cases:
|
|
if case['class'] not in by_class:
|
|
by_class[case['class']] = []
|
|
by_class[case['class']].append(case)
|
|
|
|
# Print summary
|
|
print("<table><tr><td>Class</td><td>Test Case</td><td>Failure Rate</td><td>Build Scans</td></tr>")
|
|
for full_class_name, cases in by_class.items():
|
|
print(f"<tr><td colspan=\"4\">{full_class_name}</td></tr>")
|
|
for case in cases:
|
|
method = case['method']
|
|
if method != 'N/A':
|
|
print(f"<tr><td></td><td>{method:<60}</td><td>{case['failure_rate']:.2%}</td><td>{case['total_runs']}</td></tr>")
|
|
else:
|
|
print(f"<tr><td></td><td></td><td>{case['failure_rate']:.2%}</td><td>{case['total_runs']}</td></tr>")
|
|
print("</table>")
|
|
|
|
def main():
|
|
token = None
|
|
if os.environ.get("DEVELOCITY_ACCESS_TOKEN"):
|
|
token = os.environ.get("DEVELOCITY_ACCESS_TOKEN")
|
|
else:
|
|
print("No auth token was specified. You must set DEVELOCITY_ACCESS_TOKEN to your personal access token.")
|
|
exit(1)
|
|
|
|
# Configuration
|
|
BASE_URL = "https://ge.apache.org"
|
|
PROJECT = "kafka"
|
|
QUARANTINE_THRESHOLD_DAYS = 7
|
|
MIN_FAILURE_RATE = 0.1
|
|
RECENT_FAILURE_THRESHOLD = 0.5
|
|
SUCCESS_THRESHOLD = 0.7 # For cleared tests
|
|
MIN_FLAKY_RATE = 0.2 # For flaky regressions
|
|
|
|
analyzer = TestAnalyzer(BASE_URL, token)
|
|
|
|
try:
|
|
# Get quarantined test results
|
|
quarantined_results = analyzer.get_test_results(
|
|
PROJECT,
|
|
threshold_days=QUARANTINE_THRESHOLD_DAYS,
|
|
test_type="quarantinedTest"
|
|
)
|
|
|
|
# Get regular test results for flaky regression analysis
|
|
regular_results = analyzer.get_test_results(
|
|
PROJECT,
|
|
threshold_days=7, # Last 7 days for regular tests
|
|
test_type="test"
|
|
)
|
|
|
|
# Generate reports
|
|
problematic_tests = analyzer.get_problematic_quarantined_tests(
|
|
quarantined_results,
|
|
QUARANTINE_THRESHOLD_DAYS,
|
|
MIN_FAILURE_RATE,
|
|
RECENT_FAILURE_THRESHOLD
|
|
)
|
|
|
|
flaky_regressions = analyzer.get_flaky_test_regressions(
|
|
PROJECT,
|
|
regular_results,
|
|
recent_days=7,
|
|
min_flaky_rate=MIN_FLAKY_RATE
|
|
)
|
|
|
|
cleared_tests = analyzer.get_cleared_tests(
|
|
PROJECT,
|
|
quarantined_results,
|
|
success_threshold=SUCCESS_THRESHOLD
|
|
)
|
|
|
|
# Print summary first
|
|
print(f"\n# Flaky Test Report for {datetime.now(pytz.UTC).strftime('%Y-%m-%d')}")
|
|
print(f"This report was run on {datetime.now(pytz.UTC).strftime('%Y-%m-%d %H:%M:%S')} UTC")
|
|
|
|
print_summary(problematic_tests, flaky_regressions)
|
|
|
|
# Print Flaky Test Regressions
|
|
print("\n## Flaky Test Regressions")
|
|
if not flaky_regressions:
|
|
print("No flaky test regressions found.")
|
|
else:
|
|
for test_name, details in flaky_regressions.items():
|
|
print(f"\n{test_name}")
|
|
print(f"Recent Flaky Rate: {details['recent_flaky_rate']:.2%}")
|
|
print(f"Historical Flaky Rate: {details['historical_flaky_rate']:.2%}")
|
|
print(f"\nRecent Executions (last {len(details['recent_executions'])} runs):")
|
|
for entry in sorted(details['recent_executions'], key=lambda x: x.timestamp)[-5:]:
|
|
print(f" {entry.timestamp.strftime('%Y-%m-%d %H:%M')} - {entry.outcome}")
|
|
|
|
# Print Cleared Tests
|
|
print("\n## Cleared Tests (Ready for Unquarantine)")
|
|
if not cleared_tests:
|
|
print("No tests ready to be cleared from quarantine.")
|
|
else:
|
|
# Print summary
|
|
print("<table><tr><td>Class</td><td>Test Case</td><td>Success Rate</td><td>Build Scans</td></tr>")
|
|
for test_name, details in cleared_tests.items():
|
|
print(f"<tr><td>{test_name}</td><td></td><td>{details['success_rate']:.2%}</td><td>{details['total_executions']}</td></tr>")
|
|
print("</table>")
|
|
|
|
for test_name, details in cleared_tests.items():
|
|
print(f"\n{test_name}")
|
|
print(f"Success Rate: {details['success_rate']:.2%}")
|
|
print(f"Total Executions: {details['total_executions']}")
|
|
print(f"\nRecent Executions (last {len(details['recent_executions'])} runs):")
|
|
for entry in sorted(details['recent_executions'], key=lambda x: x.timestamp):
|
|
print(f" {entry.timestamp.strftime('%Y-%m-%d %H:%M')} - {entry.outcome}")
|
|
|
|
# Print Defective Tests
|
|
print("\n## High-Priority Quarantined Tests")
|
|
if not problematic_tests:
|
|
print("No high-priority quarantined tests found.")
|
|
else:
|
|
print("These are tests which have been quarantined for several days and need attention.")
|
|
sorted_tests = sorted(
|
|
problematic_tests.items(),
|
|
key=lambda x: (x[1]['failure_rate'], x[1]['days_quarantined']),
|
|
reverse=True
|
|
)
|
|
|
|
print(f"\nFound {len(sorted_tests)} high-priority quarantined test classes:")
|
|
for full_class_name, details in sorted_tests:
|
|
class_result = details['container_result']
|
|
class_name = full_class_name.split(".")[-1]
|
|
print(f"### {class_name}")
|
|
print(f"{full_class_name} has been quarantined for {details['days_quarantined']} days")
|
|
print(f"Overall class failure: {details['failure_rate']:.2%}")
|
|
print(f"Recent class failure: {details['recent_failure_rate']:.2%}")
|
|
print("\nOverall Build Outcomes:")
|
|
print(f" Total Runs: {class_result.outcome_distribution.total}")
|
|
print(f" Failed: {class_result.outcome_distribution.failed}")
|
|
print(f" Flaky: {class_result.outcome_distribution.flaky}")
|
|
print(f" Passed: {class_result.outcome_distribution.passed}")
|
|
|
|
print("\nQuarantined Methods (Last 7 Days):")
|
|
|
|
# Sort test methods by failure rate
|
|
sorted_methods = sorted(
|
|
details['test_cases'],
|
|
key=lambda x: (x.outcome_distribution.failed + x.outcome_distribution.flaky) / x.outcome_distribution.total if x.outcome_distribution.total > 0 else 0,
|
|
reverse=True
|
|
)
|
|
|
|
for test_method in sorted_methods:
|
|
total_runs = test_method.outcome_distribution.total
|
|
if total_runs > 0:
|
|
failure_rate = (test_method.outcome_distribution.failed + test_method.outcome_distribution.flaky) / total_runs
|
|
|
|
# Extract the method name from the full test name
|
|
method_name = test_method.name.split('.')[-1]
|
|
|
|
print(f"\n → {method_name}")
|
|
print(f" Failure Rate: {failure_rate:.2%}")
|
|
print(f" Runs: {total_runs:3d} | Failed: {test_method.outcome_distribution.failed:3d} | "
|
|
f"Flaky: {test_method.outcome_distribution.flaky:3d} | "
|
|
f"Passed: {test_method.outcome_distribution.passed:3d}")
|
|
|
|
# Show test method timeline
|
|
if test_method.timeline:
|
|
print(f"\n Recent Executions (last {min(3, len(test_method.timeline))} of {len(test_method.timeline)} runs):")
|
|
print(" Date/Time (UTC) Outcome Build ID")
|
|
print(" " + "-" * 44)
|
|
for entry in sorted(test_method.timeline, key=lambda x: x.timestamp)[-3:]:
|
|
date_str = entry.timestamp.strftime('%Y-%m-%d %H:%M')
|
|
print(f" {date_str:<17} {entry.outcome:<10} {entry.build_id}")
|
|
|
|
except Exception as e:
|
|
logger.exception("Error occurred during report generation")
|
|
print(f"Error occurred: {str(e)}")
|
|
|
|
|
|
if __name__ == "__main__":
|
|
# Configure logging
|
|
logging.basicConfig(
|
|
level=logging.INFO,
|
|
format='%(asctime)s - %(levelname)s - %(message)s'
|
|
)
|
|
main()
|