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
|
#!/usr/bin/env python
# Adapted from https://github.com/mrafayaleem/kafka-jython
from __future__ import absolute_import, print_function
import argparse
import pprint
import sys
import threading
import traceback
from kafka.vendor.six.moves import range
from kafka import KafkaProducer
from test.fixtures import KafkaFixture, ZookeeperFixture
def start_brokers(n):
print('Starting {0} {1}-node cluster...'.format(KafkaFixture.kafka_version, n))
print('-> 1 Zookeeper')
zk = ZookeeperFixture.instance()
print('---> {0}:{1}'.format(zk.host, zk.port))
print()
partitions = min(n, 3)
replicas = min(n, 3)
print('-> {0} Brokers [{1} partitions / {2} replicas]'.format(n, partitions, replicas))
brokers = [
KafkaFixture.instance(i, zk, zk_chroot='',
partitions=partitions, replicas=replicas)
for i in range(n)
]
for broker in brokers:
print('---> {0}:{1}'.format(broker.host, broker.port))
print()
return brokers
class ProducerPerformance(object):
@staticmethod
def run(args):
try:
props = {}
for prop in args.producer_config:
k, v = prop.split('=')
try:
v = int(v)
except ValueError:
pass
if v == 'None':
v = None
props[k] = v
if args.brokers:
brokers = start_brokers(args.brokers)
props['bootstrap_servers'] = ['{0}:{1}'.format(broker.host, broker.port)
for broker in brokers]
print("---> bootstrap_servers={0}".format(props['bootstrap_servers']))
print()
print('-> OK!')
print()
print('Initializing producer...')
record = bytes(bytearray(args.record_size))
props['metrics_sample_window_ms'] = args.stats_interval * 1000
producer = KafkaProducer(**props)
for k, v in props.items():
print('---> {0}={1}'.format(k, v))
print('---> send {0} byte records'.format(args.record_size))
print('---> report stats every {0} secs'.format(args.stats_interval))
print('---> raw metrics? {0}'.format(args.raw_metrics))
timer_stop = threading.Event()
timer = StatsReporter(args.stats_interval, producer,
event=timer_stop,
raw_metrics=args.raw_metrics)
timer.start()
print('-> OK!')
print()
for i in range(args.num_records):
producer.send(topic=args.topic, value=record)
producer.flush()
timer_stop.set()
except Exception:
exc_info = sys.exc_info()
traceback.print_exception(*exc_info)
sys.exit(1)
class StatsReporter(threading.Thread):
def __init__(self, interval, producer, event=None, raw_metrics=False):
super(StatsReporter, self).__init__()
self.interval = interval
self.producer = producer
self.event = event
self.raw_metrics = raw_metrics
def print_stats(self):
metrics = self.producer.metrics()
if self.raw_metrics:
pprint.pprint(metrics)
else:
print('{record-send-rate} records/sec ({byte-rate} B/sec),'
' {request-latency-avg} latency,'
' {record-size-avg} record size,'
' {batch-size-avg} batch size,'
' {records-per-request-avg} records/req'
.format(**metrics['producer-metrics']))
def print_final(self):
self.print_stats()
def run(self):
while self.event and not self.event.wait(self.interval):
self.print_stats()
else:
self.print_final()
def get_args_parser():
parser = argparse.ArgumentParser(
description='This tool is used to verify the producer performance.')
parser.add_argument(
'--topic', type=str,
help='Topic name for test',
default='kafka-python-benchmark-test')
parser.add_argument(
'--num-records', type=int,
help='number of messages to produce',
default=1000000)
parser.add_argument(
'--record-size', type=int,
help='message size in bytes',
default=100)
parser.add_argument(
'--producer-config', type=str, nargs='+', default=(),
help='kafka producer related configuaration properties like '
'bootstrap_servers,client_id etc..')
parser.add_argument(
'--brokers', type=int,
help='Number of kafka brokers to start',
default=0)
parser.add_argument(
'--stats-interval', type=int,
help='Interval in seconds for stats reporting to console',
default=5)
parser.add_argument(
'--raw-metrics', action='store_true',
help='Enable this flag to print full metrics dict on each interval')
return parser
if __name__ == '__main__':
args = get_args_parser().parse_args()
ProducerPerformance.run(args)
|
