1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403
|
# Copyright DataStax, Inc.
#
# 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 time
from cassandra.connection import ConnectionShutdown
from cassandra.policies import HostFilterPolicy, RoundRobinPolicy, FallthroughRetryPolicy
import unittest
from cassandra.query import SimpleStatement
from cassandra import ConsistencyLevel, WriteTimeout, Unavailable, ReadTimeout
from cassandra.protocol import SyntaxException
from cassandra.cluster import NoHostAvailable, ExecutionProfile, EXEC_PROFILE_DEFAULT
from tests.integration import get_cluster, get_node, use_singledc, execute_until_pass, TestCluster
from greplin import scales
from tests.integration import BasicSharedKeyspaceUnitTestCaseRF3WM, BasicExistingKeyspaceUnitTestCase, local
import pprint as pp
def setup_module():
use_singledc()
@local
class MetricsTests(unittest.TestCase):
def setUp(self):
contact_point = ['127.0.0.2']
self.cluster = TestCluster(contact_points=contact_point, metrics_enabled=True,
execution_profiles=
{EXEC_PROFILE_DEFAULT:
ExecutionProfile(
load_balancing_policy=HostFilterPolicy(
RoundRobinPolicy(), lambda host: host.address in contact_point),
retry_policy=FallthroughRetryPolicy()
)
}
)
self.session = self.cluster.connect("test3rf", wait_for_all_pools=True)
def tearDown(self):
self.cluster.shutdown()
def test_connection_error(self):
"""
Trigger and ensure connection_errors are counted
Stop all node with the driver knowing about the "DOWN" states.
"""
# Test writes
for i in range(0, 100):
self.session.execute_async("INSERT INTO test (k, v) VALUES ({0}, {1})".format(i, i))
# Stop the cluster
get_cluster().stop(wait=True, gently=False)
try:
# Ensure the nodes are actually down
query = SimpleStatement("SELECT * FROM test", consistency_level=ConsistencyLevel.ALL)
# both exceptions can happen depending on when the connection has been detected as defunct
with self.assertRaises((NoHostAvailable, ConnectionShutdown)):
self.session.execute(query)
finally:
get_cluster().start(wait_for_binary_proto=True, wait_other_notice=True)
# Give some time for the cluster to come back up, for the next test
time.sleep(5)
self.assertGreater(self.cluster.metrics.stats.connection_errors, 0)
def test_write_timeout(self):
"""
Trigger and ensure write_timeouts are counted
Write a key, value pair. Pause a node without the coordinator node knowing about the "DOWN" state.
Attempt a write at cl.ALL and receive a WriteTimeout.
"""
# Test write
self.session.execute("INSERT INTO test (k, v) VALUES (1, 1)")
# Assert read
query = SimpleStatement("SELECT * FROM test WHERE k=1", consistency_level=ConsistencyLevel.ALL)
results = execute_until_pass(self.session, query)
self.assertTrue(results)
# Pause node so it shows as unreachable to coordinator
get_node(1).pause()
try:
# Test write
query = SimpleStatement("INSERT INTO test (k, v) VALUES (2, 2)", consistency_level=ConsistencyLevel.ALL)
with self.assertRaises(WriteTimeout):
self.session.execute(query, timeout=None)
self.assertEqual(1, self.cluster.metrics.stats.write_timeouts)
finally:
get_node(1).resume()
def test_read_timeout(self):
"""
Trigger and ensure read_timeouts are counted
Write a key, value pair. Pause a node without the coordinator node knowing about the "DOWN" state.
Attempt a read at cl.ALL and receive a ReadTimeout.
"""
# Test write
self.session.execute("INSERT INTO test (k, v) VALUES (1, 1)")
# Assert read
query = SimpleStatement("SELECT * FROM test WHERE k=1", consistency_level=ConsistencyLevel.ALL)
results = execute_until_pass(self.session, query)
self.assertTrue(results)
# Pause node so it shows as unreachable to coordinator
get_node(1).pause()
try:
# Test read
query = SimpleStatement("SELECT * FROM test", consistency_level=ConsistencyLevel.ALL)
with self.assertRaises(ReadTimeout):
self.session.execute(query, timeout=None)
self.assertEqual(1, self.cluster.metrics.stats.read_timeouts)
finally:
get_node(1).resume()
def test_unavailable(self):
"""
Trigger and ensure unavailables are counted
Write a key, value pair. Stop a node with the coordinator node knowing about the "DOWN" state.
Attempt an insert/read at cl.ALL and receive a Unavailable Exception.
"""
# Test write
self.session.execute("INSERT INTO test (k, v) VALUES (1, 1)")
# Assert read
query = SimpleStatement("SELECT * FROM test WHERE k=1", consistency_level=ConsistencyLevel.ALL)
results = execute_until_pass(self.session, query)
self.assertTrue(results)
# Stop node gracefully
# Sometimes this commands continues with the other nodes having not noticed
# 1 is down, and a Timeout error is returned instead of an Unavailable
get_node(1).stop(wait=True, wait_other_notice=True)
time.sleep(5)
try:
# Test write
query = SimpleStatement("INSERT INTO test (k, v) VALUES (2, 2)", consistency_level=ConsistencyLevel.ALL)
with self.assertRaises(Unavailable):
self.session.execute(query)
self.assertEqual(self.cluster.metrics.stats.unavailables, 1)
# Test write
query = SimpleStatement("SELECT * FROM test", consistency_level=ConsistencyLevel.ALL)
with self.assertRaises(Unavailable):
self.session.execute(query, timeout=None)
self.assertEqual(self.cluster.metrics.stats.unavailables, 2)
finally:
get_node(1).start(wait_other_notice=True, wait_for_binary_proto=True)
# Give some time for the cluster to come back up, for the next test
time.sleep(5)
self.cluster.shutdown()
# def test_other_error(self):
# # TODO: Bootstrapping or Overloaded cases
# pass
#
# def test_ignore(self):
# # TODO: Look for ways to generate ignores
# pass
#
# def test_retry(self):
# # TODO: Look for ways to generate retries
# pass
class MetricsNamespaceTest(BasicSharedKeyspaceUnitTestCaseRF3WM):
@local
def test_metrics_per_cluster(self):
"""
Test to validate that metrics can be scopped to invdividual clusters
@since 3.6.0
@jira_ticket PYTHON-561
@expected_result metrics should be scopped to a cluster level
@test_category metrics
"""
cluster2 = TestCluster(
metrics_enabled=True,
execution_profiles={EXEC_PROFILE_DEFAULT: ExecutionProfile(retry_policy=FallthroughRetryPolicy())}
)
cluster2.connect(self.ks_name, wait_for_all_pools=True)
self.assertEqual(len(cluster2.metadata.all_hosts()), 3)
query = SimpleStatement("SELECT * FROM {0}.{0}".format(self.ks_name), consistency_level=ConsistencyLevel.ALL)
self.session.execute(query)
# Pause node so it shows as unreachable to coordinator
get_node(1).pause()
try:
# Test write
query = SimpleStatement("INSERT INTO {0}.{0} (k, v) VALUES (2, 2)".format(self.ks_name), consistency_level=ConsistencyLevel.ALL)
with self.assertRaises(WriteTimeout):
self.session.execute(query, timeout=None)
finally:
get_node(1).resume()
# Change the scales stats_name of the cluster2
cluster2.metrics.set_stats_name('cluster2-metrics')
stats_cluster1 = self.cluster.metrics.get_stats()
stats_cluster2 = cluster2.metrics.get_stats()
# Test direct access to stats
self.assertEqual(1, self.cluster.metrics.stats.write_timeouts)
self.assertEqual(0, cluster2.metrics.stats.write_timeouts)
# Test direct access to a child stats
self.assertNotEqual(0.0, self.cluster.metrics.request_timer['mean'])
self.assertEqual(0.0, cluster2.metrics.request_timer['mean'])
# Test access via metrics.get_stats()
self.assertNotEqual(0.0, stats_cluster1['request_timer']['mean'])
self.assertEqual(0.0, stats_cluster2['request_timer']['mean'])
# Test access by stats_name
self.assertEqual(0.0, scales.getStats()['cluster2-metrics']['request_timer']['mean'])
cluster2.shutdown()
def test_duplicate_metrics_per_cluster(self):
"""
Test to validate that cluster metrics names can't overlap.
@since 3.6.0
@jira_ticket PYTHON-561
@expected_result metric names should not be allowed to be same.
@test_category metrics
"""
cluster2 = TestCluster(
metrics_enabled=True,
monitor_reporting_enabled=False,
execution_profiles={EXEC_PROFILE_DEFAULT: ExecutionProfile(retry_policy=FallthroughRetryPolicy())}
)
cluster3 = TestCluster(
metrics_enabled=True,
monitor_reporting_enabled=False,
execution_profiles={EXEC_PROFILE_DEFAULT: ExecutionProfile(retry_policy=FallthroughRetryPolicy())}
)
# Ensure duplicate metric names are not allowed
cluster2.metrics.set_stats_name("appcluster")
cluster2.metrics.set_stats_name("appcluster")
with self.assertRaises(ValueError):
cluster3.metrics.set_stats_name("appcluster")
cluster3.metrics.set_stats_name("devops")
session2 = cluster2.connect(self.ks_name, wait_for_all_pools=True)
session3 = cluster3.connect(self.ks_name, wait_for_all_pools=True)
# Basic validation that naming metrics doesn't impact their segration or accuracy
for i in range(10):
query = SimpleStatement("SELECT * FROM {0}.{0}".format(self.ks_name), consistency_level=ConsistencyLevel.ALL)
session2.execute(query)
for i in range(5):
query = SimpleStatement("SELECT * FROM {0}.{0}".format(self.ks_name), consistency_level=ConsistencyLevel.ALL)
session3.execute(query)
self.assertEqual(cluster2.metrics.get_stats()['request_timer']['count'], 10)
self.assertEqual(cluster3.metrics.get_stats()['request_timer']['count'], 5)
# Check scales to ensure they are appropriately named
self.assertTrue("appcluster" in scales._Stats.stats.keys())
self.assertTrue("devops" in scales._Stats.stats.keys())
cluster2.shutdown()
cluster3.shutdown()
class RequestAnalyzer(object):
"""
Class used to track request and error counts for a Session.
Also computes statistics on encoded request size.
"""
requests = scales.PmfStat('request size')
errors = scales.IntStat('errors')
successful = scales.IntStat("success")
# Throw exceptions when invoked.
throw_on_success = False
throw_on_fail = False
def __init__(self, session, throw_on_success=False, throw_on_fail=False):
scales.init(self, '/request')
# each instance will be registered with a session, and receive a callback for each request generated
session.add_request_init_listener(self.on_request)
self.throw_on_fail = throw_on_fail
self.throw_on_success = throw_on_success
def on_request(self, rf):
# This callback is invoked each time a request is created, on the thread creating the request.
# We can use this to count events, or add callbacks
rf.add_callbacks(self.on_success, self.on_error, callback_args=(rf,), errback_args=(rf,))
def on_success(self, _, response_future):
# future callback on a successful request; just record the size
self.requests.addValue(response_future.request_encoded_size)
self.successful += 1
if self.throw_on_success:
raise AttributeError
def on_error(self, _, response_future):
# future callback for failed; record size and increment errors
self.requests.addValue(response_future.request_encoded_size)
self.errors += 1
if self.throw_on_fail:
raise AttributeError
def remove_ra(self, session):
session.remove_request_init_listener(self.on_request)
def __str__(self):
# just extracting request count from the size stats (which are recorded on all requests)
request_sizes = dict(self.requests)
count = request_sizes.pop('count')
return "%d requests (%d errors)\nRequest size statistics:\n%s" % (count, self.errors, pp.pformat(request_sizes))
class MetricsRequestSize(BasicExistingKeyspaceUnitTestCase):
@classmethod
def setUpClass(cls):
cls.common_setup(1, keyspace_creation=False, monitor_reporting_enabled=False)
def wait_for_count(self, ra, expected_count, error=False):
for _ in range(10):
if not error:
if ra.successful is expected_count:
return True
else:
if ra.errors is expected_count:
return True
time.sleep(.01)
return False
def test_metrics_per_cluster(self):
"""
Test to validate that requests listeners.
This test creates a simple metrics based request listener to track request size, it then
check to ensure that on_success and on_error methods are invoked appropriately.
@since 3.7.0
@jira_ticket PYTHON-284
@expected_result in_error, and on_success should be invoked apropriately
@test_category metrics
"""
ra = RequestAnalyzer(self.session)
for _ in range(10):
self.session.execute("SELECT release_version FROM system.local")
for _ in range(3):
try:
self.session.execute("nonesense")
except SyntaxException:
continue
self.assertTrue(self.wait_for_count(ra, 10))
self.assertTrue(self.wait_for_count(ra, 3, error=True))
ra.remove_ra(self.session)
# Make sure a poorly coded RA doesn't cause issues
ra = RequestAnalyzer(self.session, throw_on_success=False, throw_on_fail=True)
self.session.execute("SELECT release_version FROM system.local")
ra.remove_ra(self.session)
RequestAnalyzer(self.session, throw_on_success=True)
try:
self.session.execute("nonesense")
except SyntaxException:
pass
|