File: azure_eventhub_consumer_stress_sync.py

package info (click to toggle)
python-azure 20250603%2Bgit-1
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid, trixie
  • size: 851,724 kB
  • sloc: python: 7,362,925; ansic: 804; javascript: 287; makefile: 195; sh: 145; xml: 109
file content (304 lines) | stat: -rw-r--r-- 13,368 bytes parent folder | download
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
# --------------------------------------------------------------------------------------------
# Copyright (c) Microsoft Corporation. All rights reserved.
# Licensed under the MIT License. See License.txt in the project root for license information.
# --------------------------------------------------------------------------------------------

import time
import datetime
import argparse
import threading
import os
import logging
from collections import defaultdict
from functools import partial
from dotenv import load_dotenv

from azure.identity import ClientSecretCredential, DefaultAzureCredential
from azure.eventhub.extensions.checkpointstoreblob import BlobCheckpointStore
from azure.eventhub import EventHubConsumerClient, TransportType, EventHubSharedKeyCredential

from logger import get_logger
from process_monitor import ProcessMonitor
from app_insights_metric import AzureMonitorMetric

ENV_FILE = os.environ.get("ENV_FILE")
load_dotenv(dotenv_path=ENV_FILE, override=True)


def parse_starting_position(args):
    starting_position = None
    if args.starting_offset:
        starting_position = str(args.starting_offset)
    if args.starting_sequence_number:
        starting_position = int(args.starting_sequence_number)
    if args.starting_datetime:
        starting_position = datetime.datetime.strptime(str(args.starting_datetime), "%Y-%m-%d %H:%M:%S")

    if not starting_position:
        starting_position = "-1"

    return starting_position


parser = argparse.ArgumentParser()
parser.add_argument("--link_credit", default=int(os.environ.get("LINK_CREDIT", 3000)), type=int)
parser.add_argument("--output_interval", type=float, default=int(os.environ.get("OUTPUT_INTERVAL", 5000)))
parser.add_argument(
    "--duration", help="Duration in seconds of the test", type=int, default=int(os.environ.get("DURATION", 999999999))
)
parser.add_argument(
    "--consumer_group", help="Consumer group name", default=os.environ.get("CONSUMER_GROUP", "$default")
)
parser.add_argument("--auth_timeout", help="Authorization Timeout", type=float, default=60)
parser.add_argument(
    "--starting_offset", help="Starting offset", type=str, default=os.environ.get("STARTING_OFFSET", "-1")
)
parser.add_argument("--starting_sequence_number", help="Starting sequence number", type=int)
parser.add_argument("--starting_datetime", help="Starting datetime string, should be format of YYYY-mm-dd HH:mm:ss")
parser.add_argument(
    "--partitions", help="Number of partitions. 0 means to get partitions from eventhubs", type=int, default=0
)
parser.add_argument("--owner_level", help="The owner level, or epoch, of the consumer", type=int, default=None)
parser.add_argument("--recv_partition_id", help="Receive from a specific partition if this is set", type=int)
parser.add_argument(
    "--max_batch_size",
    type=int,
    default=int(os.environ.get("MAX_BATCH_SIZE", 0)),
    help="Call EventHubConsumerClient.receive_batch() if not 0, otherwise call receive()",
)
parser.add_argument(
    "--max_wait_time",
    type=float,
    default=0,
    help="max_wait_time of EventHubConsumerClient.receive_batch() or EventHubConsumerClient.receive()",
)

parser.add_argument("--track_last_enqueued_event_properties", action="store_true")
parser.add_argument(
    "--load_balancing_interval", help="time duration in seconds between two load balance", type=float, default=10
)
parser.add_argument("--conn_str", help="EventHub connection string", default=os.environ.get("EVENT_HUB_CONN_STR"))
parser.add_argument("--azure_identity", help="Use identity", type=bool, default=False)
parser.add_argument(
    "--hostname",
    help="The fully qualified host name for the Event Hubs namespace.",
    default=os.environ.get("EVENT_HUB_HOSTNAME"),
)
parser.add_argument("--eventhub", help="Name of EventHub", default=os.environ.get("EVENT_HUB_NAME"))
parser.add_argument("--address", help="Address URI to the EventHub entity")
parser.add_argument("--sas-policy", help="Name of the shared access policy to authenticate with")
parser.add_argument("--sas-key", help="Shared access key")
parser.add_argument(
    "--transport_type", help="Transport type, 0 means AMQP, 1 means AMQP over WebSocket", type=int, default=0
)
parser.add_argument("--parallel_recv_cnt", help="Number of receive clients doing parallel receiving", type=int)
parser.add_argument("--proxy_hostname", type=str)
parser.add_argument("--proxy_port", type=str)
parser.add_argument("--proxy_username", type=str)
parser.add_argument("--proxy_password", type=str)
parser.add_argument("--aad_client_id", help="AAD client id")
parser.add_argument("--aad_secret", help="AAD secret")
parser.add_argument("--aad_tenant_id", help="AAD tenant id")
parser.add_argument("--storage_acount", default=os.environ["AZURE_STORAGE_ACCOUNT"])
parser.add_argument("--storage_conn_str", help="conn str of storage blob to store ownership and checkpoint data")
parser.add_argument("--storage_container_name", help="storage container name to store ownership and checkpoint data")
parser.add_argument("--pyamqp_logging_enable", help="pyamqp logging enable", action="store_true")
parser.add_argument("--print_console", help="print to console", action="store_true")
parser.add_argument("--log_filename", help="log file name", type=str)
parser.add_argument("--uamqp_mode", help="Flag for uamqp or pyamqp", action="store_true")
parser.add_argument(
    "--debug_level",
    help="Flag for setting a debug level, can be Info, Debug, Warning, Error or Critical",
    type=str,
    default="Error",
)
# rotate logs by default, if you want to disable it, use --no-rotating-logs flag
parser.add_argument("--no-rotating-logs", action="store_true")


args = parser.parse_args()

rotating_logs = not args.no_rotating_logs
starting_position = parse_starting_position(args)
print_console = args.print_console or (os.environ.get("PRINT_CONSOLE") == "1")
debug_level = getattr(logging, args.debug_level.upper(), logging.ERROR)


log_filename = args.log_filename if args.log_filename else "consumer_sync"
if args.transport_type == 1:
    log_filename += "_ws.log"
else:
    log_filename += ".log"
logdir = os.environ.get("DEBUG_SHARE")
logfilepath = f"{logdir}/{log_filename}"
LOGGER = get_logger(logfilepath, "stress_receive_sync", level=debug_level, print_console=print_console, rotating_logs=rotating_logs)
LOG_PER_COUNT = args.output_interval

start_time = time.perf_counter()
recv_cnt_map = defaultdict(int)
recv_cnt_iteration_map = defaultdict(int)
recv_time_map = {}

azure_metric_monitor = AzureMonitorMetric("Sync EventHubConsumerClient")


class EventHubConsumerClientTest(EventHubConsumerClient):
    def get_partition_ids(self):
        if args.partitions != 0:
            return [str(i) for i in range(args.partitions)]
        else:
            return super(EventHubConsumerClientTest, self).get_partition_ids()


def on_event_received(process_monitor, partition_context, event):
    recv_cnt_map[partition_context.partition_id] += 1 if event else 0
    if recv_cnt_map[partition_context.partition_id] % LOG_PER_COUNT == 0:
        total_time_elapsed = time.perf_counter() - start_time
        partition_previous_time = recv_time_map.get(partition_context.partition_id)
        partition_current_time = time.perf_counter()
        recv_time_map[partition_context.partition_id] = partition_current_time
        LOGGER.info(
            "Partition: %r, Total received: %r, Time elapsed: %r, Speed since start: %r/s, Current speed: %r/s",
            partition_context.partition_id,
            recv_cnt_map[partition_context.partition_id],
            total_time_elapsed,
            recv_cnt_map[partition_context.partition_id] / total_time_elapsed,
            LOG_PER_COUNT / (partition_current_time - partition_previous_time) if partition_previous_time else None,
        )
        azure_metric_monitor.record_events_cpu_memory(
            LOG_PER_COUNT, process_monitor.cpu_usage_percent, process_monitor.memory_usage_percent
        )
        partition_context.update_checkpoint(event)


def on_event_batch_received(process_monitor, partition_context, event_batch):
    recv_cnt_map[partition_context.partition_id] += len(event_batch)
    recv_cnt_iteration_map[partition_context.partition_id] += len(event_batch)
    if recv_cnt_iteration_map[partition_context.partition_id] > LOG_PER_COUNT:
        total_time_elapsed = time.perf_counter() - start_time
        partition_previous_time = recv_time_map.get(partition_context.partition_id)
        partition_current_time = time.perf_counter()
        recv_time_map[partition_context.partition_id] = partition_current_time
        LOGGER.info(
            "Partition: %r, Total received: %r, Time elapsed: %r, Speed since start: %r/s, Current speed: %r/s",
            partition_context.partition_id,
            recv_cnt_map[partition_context.partition_id],
            total_time_elapsed,
            recv_cnt_map[partition_context.partition_id] / total_time_elapsed,
            (
                recv_cnt_iteration_map[partition_context.partition_id]
                / (partition_current_time - partition_previous_time)
                if partition_previous_time
                else None
            ),
        )
        recv_cnt_iteration_map[partition_context.partition_id] = 0
        azure_metric_monitor.record_events_cpu_memory(
            LOG_PER_COUNT, process_monitor.cpu_usage_percent, process_monitor.memory_usage_percent
        )
        partition_context.update_checkpoint()


def on_error(partition_context, exception):
    azure_metric_monitor.record_error(exception, extra="partition: {}".format(partition_context.partition_id))


def create_client(args):
    try:
        checkpoint_store = BlobCheckpointStore(
            args.storage_account, DefaultAzureCredential(), args.storage_container_name
        )
    except:
        checkpoint_store = None

    transport_type = TransportType.Amqp if args.transport_type == 0 else TransportType.AmqpOverWebsocket
    http_proxy = None
    if args.proxy_hostname:
        http_proxy = {
            "proxy_hostname": args.proxy_hostname,
            "proxy_port": args.proxy_port,
            "username": args.proxy_username,
            "password": args.proxy_password,
        }
    client = EventHubConsumerClientTest(
        fully_qualified_namespace=args.hostname,
        eventhub_name=args.eventhub,
        consumer_group=args.consumer_group,
        credential=DefaultAzureCredential(),
        checkpoint_store=checkpoint_store,
        load_balancing_interval=args.load_balancing_interval,
        auth_timeout=args.auth_timeout,
        http_proxy=http_proxy,
        transport_type=transport_type,
        logging_enable=args.pyamqp_logging_enable,
        uamqp_transport=args.uamqp_mode,
    )
    return client


def run(args):

    with ProcessMonitor(
        "monitor_{}".format(args.log_filename), "consumer_stress_sync", print_console=print_console
    ) as process_monitor:
        kwargs_dict = {
            "prefetch": args.link_credit,
            "partition_id": str(args.recv_partition_id) if args.recv_partition_id else None,
            "track_last_enqueued_event_properties": args.track_last_enqueued_event_properties,
            "starting_position": starting_position,
            "owner_level": args.owner_level,
            "on_error": on_error,
        }
        if args.max_batch_size:
            kwargs_dict["max_batch_size"] = args.max_batch_size
        if args.max_wait_time:
            kwargs_dict["max_wait_time"] = args.max_wait_time

        on_event_received_with_process_monitor = partial(on_event_received, process_monitor)
        on_event_batch_received_with_process_monitor = partial(on_event_batch_received, process_monitor)

        if args.parallel_recv_cnt and args.parallel_recv_cnt > 1:
            clients = [create_client(args) for _ in range(args.parallel_recv_cnt)]
            threads = [
                threading.Thread(
                    target=clients[i].receive_batch if args.max_batch_size else clients[i].receive,
                    args=(
                        (
                            on_event_batch_received_with_process_monitor
                            if args.max_batch_size
                            else on_event_received_with_process_monitor
                        ),
                    ),
                    kwargs=kwargs_dict,
                    daemon=True,
                )
                for i in range(args.parallel_recv_cnt)
            ]
        else:
            clients = [create_client(args)]
            threads = [
                threading.Thread(
                    target=clients[0].receive_batch if args.max_batch_size else clients[0].receive,
                    args=(
                        (
                            on_event_batch_received_with_process_monitor
                            if args.max_batch_size
                            else on_event_received_with_process_monitor
                        ),
                    ),
                    kwargs=kwargs_dict,
                    daemon=True,
                )
            ]

        for thread in threads:
            thread.start()
        time.sleep(args.duration)
        for client in clients:
            client.close()
        for thread in threads:
            thread.join()


if __name__ == "__main__":
    run(args)