karafka 2.0.33 → 2.0.34
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- checksums.yaml.gz.sig +1 -1
- data/CHANGELOG.md +9 -1
- data/Gemfile.lock +4 -4
- data/lib/karafka/embedded.rb +4 -1
- data/lib/karafka/instrumentation/vendors/datadog/listener.rb +4 -247
- data/lib/karafka/instrumentation/vendors/datadog/metrics_listener.rb +259 -0
- data/lib/karafka/version.rb +1 -1
- data.tar.gz.sig +0 -0
- metadata +3 -2
- metadata.gz.sig +0 -0
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 36d890d825aaeaee5349dcc653d888da3a023c01a837864544a905db977569c4
|
4
|
+
data.tar.gz: be442485812a05a030bab33da31a8e2fda684add8c4d59a0af78f517bb2519bd
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: d92be137485c436c1ed02435669785422e6e4da194ab19b97dca31f70b530fe7e0ae4e6b0c7c54895dbceca2d8fe4ef1bdcabdb287affe06e377942062777979
|
7
|
+
data.tar.gz: 0525b652373088a7a692134a6a4c89487e495e01934504d256a60ae7805c92f65a308d38c8d75ad806f39f5128d2d00865e10b1b4ae326c7dd60e7594c68558e
|
checksums.yaml.gz.sig
CHANGED
@@ -1 +1 @@
|
|
1
|
-
|
1
|
+
Ԫ����D�����jO����dl�0m��I���t�uNv�b��2z_����viGzW:dA|��A����՛<u��i�R�8�LW��[m�ڳ�3�y΄Py{#�)�����j�z.�::�z`���ו�,b�ֵ�:w��H-���/���Ӷ2�B�s��r6c
|
data/CHANGELOG.md
CHANGED
@@ -1,5 +1,13 @@
|
|
1
1
|
# Karafka framework changelog
|
2
2
|
|
3
|
+
## 2.0.34 (2023-03-04)
|
4
|
+
- [Improvement] Attach an `embedded` tag to Karafka processes started using the embedded API.
|
5
|
+
- [Change] Renamed `Datadog::Listener` to `Datadog::MetricsListener` for consistency. (#1124)
|
6
|
+
|
7
|
+
### Upgrade notes
|
8
|
+
|
9
|
+
1. Replace `Datadog::Listener` references to `Datadog::MetricsListener`.
|
10
|
+
|
3
11
|
## 2.0.33 (2023-02-24)
|
4
12
|
- **[Feature]** Support `perform_all_later` in ActiveJob adapter for Rails `7.1+`
|
5
13
|
- **[Feature]** Introduce ability to assign and re-assign tags in consumer instances. This can be used for extra instrumentation that is context aware.
|
@@ -99,7 +107,7 @@ class KarafkaApp < Karafka::App
|
|
99
107
|
- [Improvement] Expand `LoggerListener` with `client.resume` notification.
|
100
108
|
- [Improvement] Replace random anonymous subscription groups ids with stable once.
|
101
109
|
- [Improvement] Add `consumer.consume`, `consumer.revoke` and `consumer.shutting_down` notification events and move the revocation logic calling to strategies.
|
102
|
-
- [Change] Rename job queue statistics `processing` key to `busy`. No changes needed because naming in the DataDog listener stays the same.
|
110
|
+
- [Change] Rename job queue statistics `processing` key to `busy`. No changes needed because naming in the DataDog listener stays the same.
|
103
111
|
- [Fix] Fix proctitle listener state changes reporting on new states.
|
104
112
|
- [Fix] Make sure all files descriptors are closed in the integration specs.
|
105
113
|
- [Fix] Fix a case where empty subscription groups could leak into the execution flow.
|
data/Gemfile.lock
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
PATH
|
2
2
|
remote: .
|
3
3
|
specs:
|
4
|
-
karafka (2.0.
|
4
|
+
karafka (2.0.34)
|
5
5
|
karafka-core (>= 2.0.12, < 3.0.0)
|
6
6
|
thor (>= 0.20)
|
7
7
|
waterdrop (>= 2.4.10, < 3.0.0)
|
@@ -19,7 +19,7 @@ GEM
|
|
19
19
|
minitest (>= 5.1)
|
20
20
|
tzinfo (~> 2.0)
|
21
21
|
byebug (11.1.3)
|
22
|
-
concurrent-ruby (1.2.
|
22
|
+
concurrent-ruby (1.2.2)
|
23
23
|
diff-lcs (1.5.0)
|
24
24
|
docile (1.4.0)
|
25
25
|
factory_bot (6.2.1)
|
@@ -61,8 +61,8 @@ GEM
|
|
61
61
|
thor (1.2.1)
|
62
62
|
tzinfo (2.0.6)
|
63
63
|
concurrent-ruby (~> 1.0)
|
64
|
-
waterdrop (2.4.
|
65
|
-
karafka-core (>= 2.0.
|
64
|
+
waterdrop (2.4.11)
|
65
|
+
karafka-core (>= 2.0.12, < 3.0.0)
|
66
66
|
zeitwerk (~> 2.3)
|
67
67
|
zeitwerk (2.6.7)
|
68
68
|
|
data/lib/karafka/embedded.rb
CHANGED
@@ -7,7 +7,10 @@ module Karafka
|
|
7
7
|
# Starts Karafka without supervision and without ownership of signals in a background thread
|
8
8
|
# so it won't interrupt other things running
|
9
9
|
def start
|
10
|
-
Thread.new
|
10
|
+
Thread.new do
|
11
|
+
Karafka::Process.tags.add(:execution_mode, 'embedded')
|
12
|
+
Karafka::Server.start
|
13
|
+
end
|
11
14
|
end
|
12
15
|
|
13
16
|
# Stops Karafka upon any event
|
@@ -1,258 +1,15 @@
|
|
1
1
|
# frozen_string_literal: true
|
2
2
|
|
3
|
+
require_relative 'metrics_listener'
|
4
|
+
|
3
5
|
module Karafka
|
4
6
|
module Instrumentation
|
5
7
|
# Namespace for vendor specific instrumentation
|
6
8
|
module Vendors
|
7
9
|
# Datadog specific instrumentation
|
8
10
|
module Datadog
|
9
|
-
#
|
10
|
-
|
11
|
-
#
|
12
|
-
# @note You need to setup the `dogstatsd-ruby` client and assign it
|
13
|
-
class Listener
|
14
|
-
include ::Karafka::Core::Configurable
|
15
|
-
extend Forwardable
|
16
|
-
|
17
|
-
def_delegators :config, :client, :rd_kafka_metrics, :namespace, :default_tags
|
18
|
-
|
19
|
-
# Value object for storing a single rdkafka metric publishing details
|
20
|
-
RdKafkaMetric = Struct.new(:type, :scope, :name, :key_location)
|
21
|
-
|
22
|
-
# Namespace under which the DD metrics should be published
|
23
|
-
setting :namespace, default: 'karafka'
|
24
|
-
|
25
|
-
# Datadog client that we should use to publish the metrics
|
26
|
-
setting :client
|
27
|
-
|
28
|
-
# Default tags we want to publish (for example hostname)
|
29
|
-
# Format as followed (example for hostname): `["host:#{Socket.gethostname}"]`
|
30
|
-
setting :default_tags, default: []
|
31
|
-
|
32
|
-
# All the rdkafka metrics we want to publish
|
33
|
-
#
|
34
|
-
# By default we publish quite a lot so this can be tuned
|
35
|
-
# Note, that the once with `_d` come from Karafka, not rdkafka or Kafka
|
36
|
-
setting :rd_kafka_metrics, default: [
|
37
|
-
# Client metrics
|
38
|
-
RdKafkaMetric.new(:count, :root, 'messages.consumed', 'rxmsgs_d'),
|
39
|
-
RdKafkaMetric.new(:count, :root, 'messages.consumed.bytes', 'rxmsg_bytes'),
|
40
|
-
|
41
|
-
# Broker metrics
|
42
|
-
RdKafkaMetric.new(:count, :brokers, 'consume.attempts', 'txretries_d'),
|
43
|
-
RdKafkaMetric.new(:count, :brokers, 'consume.errors', 'txerrs_d'),
|
44
|
-
RdKafkaMetric.new(:count, :brokers, 'receive.errors', 'rxerrs_d'),
|
45
|
-
RdKafkaMetric.new(:count, :brokers, 'connection.connects', 'connects_d'),
|
46
|
-
RdKafkaMetric.new(:count, :brokers, 'connection.disconnects', 'disconnects_d'),
|
47
|
-
RdKafkaMetric.new(:gauge, :brokers, 'network.latency.avg', %w[rtt avg]),
|
48
|
-
RdKafkaMetric.new(:gauge, :brokers, 'network.latency.p95', %w[rtt p95]),
|
49
|
-
RdKafkaMetric.new(:gauge, :brokers, 'network.latency.p99', %w[rtt p99]),
|
50
|
-
|
51
|
-
# Topics metrics
|
52
|
-
RdKafkaMetric.new(:gauge, :topics, 'consumer.lags', 'consumer_lag_stored'),
|
53
|
-
RdKafkaMetric.new(:gauge, :topics, 'consumer.lags_delta', 'consumer_lag_stored_d')
|
54
|
-
].freeze
|
55
|
-
|
56
|
-
configure
|
57
|
-
|
58
|
-
# @param block [Proc] configuration block
|
59
|
-
def initialize(&block)
|
60
|
-
configure
|
61
|
-
setup(&block) if block
|
62
|
-
end
|
63
|
-
|
64
|
-
# @param block [Proc] configuration block
|
65
|
-
# @note We define this alias to be consistent with `WaterDrop#setup`
|
66
|
-
def setup(&block)
|
67
|
-
configure(&block)
|
68
|
-
end
|
69
|
-
|
70
|
-
# Hooks up to WaterDrop instrumentation for emitted statistics
|
71
|
-
#
|
72
|
-
# @param event [Karafka::Core::Monitoring::Event]
|
73
|
-
def on_statistics_emitted(event)
|
74
|
-
statistics = event[:statistics]
|
75
|
-
consumer_group_id = event[:consumer_group_id]
|
76
|
-
|
77
|
-
base_tags = default_tags + ["consumer_group:#{consumer_group_id}"]
|
78
|
-
|
79
|
-
rd_kafka_metrics.each do |metric|
|
80
|
-
report_metric(metric, statistics, base_tags)
|
81
|
-
end
|
82
|
-
end
|
83
|
-
|
84
|
-
# Increases the errors count by 1
|
85
|
-
#
|
86
|
-
# @param event [Karafka::Core::Monitoring::Event]
|
87
|
-
def on_error_occurred(event)
|
88
|
-
extra_tags = ["type:#{event[:type]}"]
|
89
|
-
|
90
|
-
if event.payload[:caller].respond_to?(:messages)
|
91
|
-
extra_tags += consumer_tags(event.payload[:caller])
|
92
|
-
end
|
93
|
-
|
94
|
-
count('error_occurred', 1, tags: default_tags + extra_tags)
|
95
|
-
end
|
96
|
-
|
97
|
-
# Reports how many messages we've polled and how much time did we spend on it
|
98
|
-
#
|
99
|
-
# @param event [Karafka::Core::Monitoring::Event]
|
100
|
-
def on_connection_listener_fetch_loop_received(event)
|
101
|
-
time_taken = event[:time]
|
102
|
-
messages_count = event[:messages_buffer].size
|
103
|
-
|
104
|
-
consumer_group_id = event[:subscription_group].consumer_group_id
|
105
|
-
|
106
|
-
extra_tags = ["consumer_group:#{consumer_group_id}"]
|
107
|
-
|
108
|
-
histogram('listener.polling.time_taken', time_taken, tags: default_tags + extra_tags)
|
109
|
-
histogram('listener.polling.messages', messages_count, tags: default_tags + extra_tags)
|
110
|
-
end
|
111
|
-
|
112
|
-
# Here we report majority of things related to processing as we have access to the
|
113
|
-
# consumer
|
114
|
-
# @param event [Karafka::Core::Monitoring::Event]
|
115
|
-
def on_consumer_consumed(event)
|
116
|
-
consumer = event.payload[:caller]
|
117
|
-
messages = consumer.messages
|
118
|
-
metadata = messages.metadata
|
119
|
-
|
120
|
-
tags = default_tags + consumer_tags(consumer)
|
121
|
-
|
122
|
-
count('consumer.messages', messages.count, tags: tags)
|
123
|
-
count('consumer.batches', 1, tags: tags)
|
124
|
-
gauge('consumer.offset', metadata.last_offset, tags: tags)
|
125
|
-
histogram('consumer.consumed.time_taken', event[:time], tags: tags)
|
126
|
-
histogram('consumer.batch_size', messages.count, tags: tags)
|
127
|
-
histogram('consumer.processing_lag', metadata.processing_lag, tags: tags)
|
128
|
-
histogram('consumer.consumption_lag', metadata.consumption_lag, tags: tags)
|
129
|
-
end
|
130
|
-
|
131
|
-
# @param event [Karafka::Core::Monitoring::Event]
|
132
|
-
def on_consumer_revoked(event)
|
133
|
-
tags = default_tags + consumer_tags(event.payload[:caller])
|
134
|
-
|
135
|
-
count('consumer.revoked', 1, tags: tags)
|
136
|
-
end
|
137
|
-
|
138
|
-
# @param event [Karafka::Core::Monitoring::Event]
|
139
|
-
def on_consumer_shutdown(event)
|
140
|
-
tags = default_tags + consumer_tags(event.payload[:caller])
|
141
|
-
|
142
|
-
count('consumer.shutdown', 1, tags: tags)
|
143
|
-
end
|
144
|
-
|
145
|
-
# Worker related metrics
|
146
|
-
# @param event [Karafka::Core::Monitoring::Event]
|
147
|
-
def on_worker_process(event)
|
148
|
-
jq_stats = event[:jobs_queue].statistics
|
149
|
-
|
150
|
-
gauge('worker.total_threads', Karafka::App.config.concurrency, tags: default_tags)
|
151
|
-
histogram('worker.processing', jq_stats[:busy], tags: default_tags)
|
152
|
-
histogram('worker.enqueued_jobs', jq_stats[:enqueued], tags: default_tags)
|
153
|
-
end
|
154
|
-
|
155
|
-
# We report this metric before and after processing for higher accuracy
|
156
|
-
# Without this, the utilization would not be fully reflected
|
157
|
-
# @param event [Karafka::Core::Monitoring::Event]
|
158
|
-
def on_worker_processed(event)
|
159
|
-
jq_stats = event[:jobs_queue].statistics
|
160
|
-
|
161
|
-
histogram('worker.processing', jq_stats[:busy], tags: default_tags)
|
162
|
-
end
|
163
|
-
|
164
|
-
private
|
165
|
-
|
166
|
-
%i[
|
167
|
-
count
|
168
|
-
gauge
|
169
|
-
histogram
|
170
|
-
increment
|
171
|
-
decrement
|
172
|
-
].each do |metric_type|
|
173
|
-
class_eval <<~METHODS, __FILE__, __LINE__ + 1
|
174
|
-
def #{metric_type}(key, *args)
|
175
|
-
client.#{metric_type}(
|
176
|
-
namespaced_metric(key),
|
177
|
-
*args
|
178
|
-
)
|
179
|
-
end
|
180
|
-
METHODS
|
181
|
-
end
|
182
|
-
|
183
|
-
# Wraps metric name in listener's namespace
|
184
|
-
# @param metric_name [String] RdKafkaMetric name
|
185
|
-
# @return [String]
|
186
|
-
def namespaced_metric(metric_name)
|
187
|
-
"#{namespace}.#{metric_name}"
|
188
|
-
end
|
189
|
-
|
190
|
-
# Reports a given metric statistics to Datadog
|
191
|
-
# @param metric [RdKafkaMetric] metric value object
|
192
|
-
# @param statistics [Hash] hash with all the statistics emitted
|
193
|
-
# @param base_tags [Array<String>] base tags we want to start with
|
194
|
-
def report_metric(metric, statistics, base_tags)
|
195
|
-
case metric.scope
|
196
|
-
when :root
|
197
|
-
public_send(
|
198
|
-
metric.type,
|
199
|
-
metric.name,
|
200
|
-
statistics.fetch(*metric.key_location),
|
201
|
-
tags: base_tags
|
202
|
-
)
|
203
|
-
when :brokers
|
204
|
-
statistics.fetch('brokers').each_value do |broker_statistics|
|
205
|
-
# Skip bootstrap nodes
|
206
|
-
# Bootstrap nodes have nodeid -1, other nodes have positive
|
207
|
-
# node ids
|
208
|
-
next if broker_statistics['nodeid'] == -1
|
209
|
-
|
210
|
-
public_send(
|
211
|
-
metric.type,
|
212
|
-
metric.name,
|
213
|
-
broker_statistics.dig(*metric.key_location),
|
214
|
-
tags: base_tags + ["broker:#{broker_statistics['nodename']}"]
|
215
|
-
)
|
216
|
-
end
|
217
|
-
when :topics
|
218
|
-
statistics.fetch('topics').each do |topic_name, topic_values|
|
219
|
-
topic_values['partitions'].each do |partition_name, partition_statistics|
|
220
|
-
next if partition_name == '-1'
|
221
|
-
# Skip until lag info is available
|
222
|
-
next if partition_statistics['consumer_lag'] == -1
|
223
|
-
|
224
|
-
public_send(
|
225
|
-
metric.type,
|
226
|
-
metric.name,
|
227
|
-
partition_statistics.dig(*metric.key_location),
|
228
|
-
tags: base_tags + [
|
229
|
-
"topic:#{topic_name}",
|
230
|
-
"partition:#{partition_name}"
|
231
|
-
]
|
232
|
-
)
|
233
|
-
end
|
234
|
-
end
|
235
|
-
else
|
236
|
-
raise ArgumentError, metric.scope
|
237
|
-
end
|
238
|
-
end
|
239
|
-
|
240
|
-
# Builds basic per consumer tags for publication
|
241
|
-
#
|
242
|
-
# @param consumer [Karafka::BaseConsumer]
|
243
|
-
# @return [Array<String>]
|
244
|
-
def consumer_tags(consumer)
|
245
|
-
messages = consumer.messages
|
246
|
-
metadata = messages.metadata
|
247
|
-
consumer_group_id = consumer.topic.consumer_group.id
|
248
|
-
|
249
|
-
[
|
250
|
-
"topic:#{metadata.topic}",
|
251
|
-
"partition:#{metadata.partition}",
|
252
|
-
"consumer_group:#{consumer_group_id}"
|
253
|
-
]
|
254
|
-
end
|
255
|
-
end
|
11
|
+
# Alias to keep backwards compatibility
|
12
|
+
Listener = MetricsListener
|
256
13
|
end
|
257
14
|
end
|
258
15
|
end
|
@@ -0,0 +1,259 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Karafka
|
4
|
+
module Instrumentation
|
5
|
+
# Namespace for vendor specific instrumentation
|
6
|
+
module Vendors
|
7
|
+
# Datadog specific instrumentation
|
8
|
+
module Datadog
|
9
|
+
# Listener that can be used to subscribe to Karafka to receive stats via StatsD
|
10
|
+
# and/or Datadog
|
11
|
+
#
|
12
|
+
# @note You need to setup the `dogstatsd-ruby` client and assign it
|
13
|
+
class MetricsListener
|
14
|
+
include ::Karafka::Core::Configurable
|
15
|
+
extend Forwardable
|
16
|
+
|
17
|
+
def_delegators :config, :client, :rd_kafka_metrics, :namespace, :default_tags
|
18
|
+
|
19
|
+
# Value object for storing a single rdkafka metric publishing details
|
20
|
+
RdKafkaMetric = Struct.new(:type, :scope, :name, :key_location)
|
21
|
+
|
22
|
+
# Namespace under which the DD metrics should be published
|
23
|
+
setting :namespace, default: 'karafka'
|
24
|
+
|
25
|
+
# Datadog client that we should use to publish the metrics
|
26
|
+
setting :client
|
27
|
+
|
28
|
+
# Default tags we want to publish (for example hostname)
|
29
|
+
# Format as followed (example for hostname): `["host:#{Socket.gethostname}"]`
|
30
|
+
setting :default_tags, default: []
|
31
|
+
|
32
|
+
# All the rdkafka metrics we want to publish
|
33
|
+
#
|
34
|
+
# By default we publish quite a lot so this can be tuned
|
35
|
+
# Note, that the once with `_d` come from Karafka, not rdkafka or Kafka
|
36
|
+
setting :rd_kafka_metrics, default: [
|
37
|
+
# Client metrics
|
38
|
+
RdKafkaMetric.new(:count, :root, 'messages.consumed', 'rxmsgs_d'),
|
39
|
+
RdKafkaMetric.new(:count, :root, 'messages.consumed.bytes', 'rxmsg_bytes'),
|
40
|
+
|
41
|
+
# Broker metrics
|
42
|
+
RdKafkaMetric.new(:count, :brokers, 'consume.attempts', 'txretries_d'),
|
43
|
+
RdKafkaMetric.new(:count, :brokers, 'consume.errors', 'txerrs_d'),
|
44
|
+
RdKafkaMetric.new(:count, :brokers, 'receive.errors', 'rxerrs_d'),
|
45
|
+
RdKafkaMetric.new(:count, :brokers, 'connection.connects', 'connects_d'),
|
46
|
+
RdKafkaMetric.new(:count, :brokers, 'connection.disconnects', 'disconnects_d'),
|
47
|
+
RdKafkaMetric.new(:gauge, :brokers, 'network.latency.avg', %w[rtt avg]),
|
48
|
+
RdKafkaMetric.new(:gauge, :brokers, 'network.latency.p95', %w[rtt p95]),
|
49
|
+
RdKafkaMetric.new(:gauge, :brokers, 'network.latency.p99', %w[rtt p99]),
|
50
|
+
|
51
|
+
# Topics metrics
|
52
|
+
RdKafkaMetric.new(:gauge, :topics, 'consumer.lags', 'consumer_lag_stored'),
|
53
|
+
RdKafkaMetric.new(:gauge, :topics, 'consumer.lags_delta', 'consumer_lag_stored_d')
|
54
|
+
].freeze
|
55
|
+
|
56
|
+
configure
|
57
|
+
|
58
|
+
# @param block [Proc] configuration block
|
59
|
+
def initialize(&block)
|
60
|
+
configure
|
61
|
+
setup(&block) if block
|
62
|
+
end
|
63
|
+
|
64
|
+
# @param block [Proc] configuration block
|
65
|
+
# @note We define this alias to be consistent with `WaterDrop#setup`
|
66
|
+
def setup(&block)
|
67
|
+
configure(&block)
|
68
|
+
end
|
69
|
+
|
70
|
+
# Hooks up to WaterDrop instrumentation for emitted statistics
|
71
|
+
#
|
72
|
+
# @param event [Karafka::Core::Monitoring::Event]
|
73
|
+
def on_statistics_emitted(event)
|
74
|
+
statistics = event[:statistics]
|
75
|
+
consumer_group_id = event[:consumer_group_id]
|
76
|
+
|
77
|
+
base_tags = default_tags + ["consumer_group:#{consumer_group_id}"]
|
78
|
+
|
79
|
+
rd_kafka_metrics.each do |metric|
|
80
|
+
report_metric(metric, statistics, base_tags)
|
81
|
+
end
|
82
|
+
end
|
83
|
+
|
84
|
+
# Increases the errors count by 1
|
85
|
+
#
|
86
|
+
# @param event [Karafka::Core::Monitoring::Event]
|
87
|
+
def on_error_occurred(event)
|
88
|
+
extra_tags = ["type:#{event[:type]}"]
|
89
|
+
|
90
|
+
if event.payload[:caller].respond_to?(:messages)
|
91
|
+
extra_tags += consumer_tags(event.payload[:caller])
|
92
|
+
end
|
93
|
+
|
94
|
+
count('error_occurred', 1, tags: default_tags + extra_tags)
|
95
|
+
end
|
96
|
+
|
97
|
+
# Reports how many messages we've polled and how much time did we spend on it
|
98
|
+
#
|
99
|
+
# @param event [Karafka::Core::Monitoring::Event]
|
100
|
+
def on_connection_listener_fetch_loop_received(event)
|
101
|
+
time_taken = event[:time]
|
102
|
+
messages_count = event[:messages_buffer].size
|
103
|
+
|
104
|
+
consumer_group_id = event[:subscription_group].consumer_group_id
|
105
|
+
|
106
|
+
extra_tags = ["consumer_group:#{consumer_group_id}"]
|
107
|
+
|
108
|
+
histogram('listener.polling.time_taken', time_taken, tags: default_tags + extra_tags)
|
109
|
+
histogram('listener.polling.messages', messages_count, tags: default_tags + extra_tags)
|
110
|
+
end
|
111
|
+
|
112
|
+
# Here we report majority of things related to processing as we have access to the
|
113
|
+
# consumer
|
114
|
+
# @param event [Karafka::Core::Monitoring::Event]
|
115
|
+
def on_consumer_consumed(event)
|
116
|
+
consumer = event.payload[:caller]
|
117
|
+
messages = consumer.messages
|
118
|
+
metadata = messages.metadata
|
119
|
+
|
120
|
+
tags = default_tags + consumer_tags(consumer)
|
121
|
+
|
122
|
+
count('consumer.messages', messages.count, tags: tags)
|
123
|
+
count('consumer.batches', 1, tags: tags)
|
124
|
+
gauge('consumer.offset', metadata.last_offset, tags: tags)
|
125
|
+
histogram('consumer.consumed.time_taken', event[:time], tags: tags)
|
126
|
+
histogram('consumer.batch_size', messages.count, tags: tags)
|
127
|
+
histogram('consumer.processing_lag', metadata.processing_lag, tags: tags)
|
128
|
+
histogram('consumer.consumption_lag', metadata.consumption_lag, tags: tags)
|
129
|
+
end
|
130
|
+
|
131
|
+
# @param event [Karafka::Core::Monitoring::Event]
|
132
|
+
def on_consumer_revoked(event)
|
133
|
+
tags = default_tags + consumer_tags(event.payload[:caller])
|
134
|
+
|
135
|
+
count('consumer.revoked', 1, tags: tags)
|
136
|
+
end
|
137
|
+
|
138
|
+
# @param event [Karafka::Core::Monitoring::Event]
|
139
|
+
def on_consumer_shutdown(event)
|
140
|
+
tags = default_tags + consumer_tags(event.payload[:caller])
|
141
|
+
|
142
|
+
count('consumer.shutdown', 1, tags: tags)
|
143
|
+
end
|
144
|
+
|
145
|
+
# Worker related metrics
|
146
|
+
# @param event [Karafka::Core::Monitoring::Event]
|
147
|
+
def on_worker_process(event)
|
148
|
+
jq_stats = event[:jobs_queue].statistics
|
149
|
+
|
150
|
+
gauge('worker.total_threads', Karafka::App.config.concurrency, tags: default_tags)
|
151
|
+
histogram('worker.processing', jq_stats[:busy], tags: default_tags)
|
152
|
+
histogram('worker.enqueued_jobs', jq_stats[:enqueued], tags: default_tags)
|
153
|
+
end
|
154
|
+
|
155
|
+
# We report this metric before and after processing for higher accuracy
|
156
|
+
# Without this, the utilization would not be fully reflected
|
157
|
+
# @param event [Karafka::Core::Monitoring::Event]
|
158
|
+
def on_worker_processed(event)
|
159
|
+
jq_stats = event[:jobs_queue].statistics
|
160
|
+
|
161
|
+
histogram('worker.processing', jq_stats[:busy], tags: default_tags)
|
162
|
+
end
|
163
|
+
|
164
|
+
private
|
165
|
+
|
166
|
+
%i[
|
167
|
+
count
|
168
|
+
gauge
|
169
|
+
histogram
|
170
|
+
increment
|
171
|
+
decrement
|
172
|
+
].each do |metric_type|
|
173
|
+
class_eval <<~METHODS, __FILE__, __LINE__ + 1
|
174
|
+
def #{metric_type}(key, *args)
|
175
|
+
client.#{metric_type}(
|
176
|
+
namespaced_metric(key),
|
177
|
+
*args
|
178
|
+
)
|
179
|
+
end
|
180
|
+
METHODS
|
181
|
+
end
|
182
|
+
|
183
|
+
# Wraps metric name in listener's namespace
|
184
|
+
# @param metric_name [String] RdKafkaMetric name
|
185
|
+
# @return [String]
|
186
|
+
def namespaced_metric(metric_name)
|
187
|
+
"#{namespace}.#{metric_name}"
|
188
|
+
end
|
189
|
+
|
190
|
+
# Reports a given metric statistics to Datadog
|
191
|
+
# @param metric [RdKafkaMetric] metric value object
|
192
|
+
# @param statistics [Hash] hash with all the statistics emitted
|
193
|
+
# @param base_tags [Array<String>] base tags we want to start with
|
194
|
+
def report_metric(metric, statistics, base_tags)
|
195
|
+
case metric.scope
|
196
|
+
when :root
|
197
|
+
public_send(
|
198
|
+
metric.type,
|
199
|
+
metric.name,
|
200
|
+
statistics.fetch(*metric.key_location),
|
201
|
+
tags: base_tags
|
202
|
+
)
|
203
|
+
when :brokers
|
204
|
+
statistics.fetch('brokers').each_value do |broker_statistics|
|
205
|
+
# Skip bootstrap nodes
|
206
|
+
# Bootstrap nodes have nodeid -1, other nodes have positive
|
207
|
+
# node ids
|
208
|
+
next if broker_statistics['nodeid'] == -1
|
209
|
+
|
210
|
+
public_send(
|
211
|
+
metric.type,
|
212
|
+
metric.name,
|
213
|
+
broker_statistics.dig(*metric.key_location),
|
214
|
+
tags: base_tags + ["broker:#{broker_statistics['nodename']}"]
|
215
|
+
)
|
216
|
+
end
|
217
|
+
when :topics
|
218
|
+
statistics.fetch('topics').each do |topic_name, topic_values|
|
219
|
+
topic_values['partitions'].each do |partition_name, partition_statistics|
|
220
|
+
next if partition_name == '-1'
|
221
|
+
# Skip until lag info is available
|
222
|
+
next if partition_statistics['consumer_lag'] == -1
|
223
|
+
|
224
|
+
public_send(
|
225
|
+
metric.type,
|
226
|
+
metric.name,
|
227
|
+
partition_statistics.dig(*metric.key_location),
|
228
|
+
tags: base_tags + [
|
229
|
+
"topic:#{topic_name}",
|
230
|
+
"partition:#{partition_name}"
|
231
|
+
]
|
232
|
+
)
|
233
|
+
end
|
234
|
+
end
|
235
|
+
else
|
236
|
+
raise ArgumentError, metric.scope
|
237
|
+
end
|
238
|
+
end
|
239
|
+
|
240
|
+
# Builds basic per consumer tags for publication
|
241
|
+
#
|
242
|
+
# @param consumer [Karafka::BaseConsumer]
|
243
|
+
# @return [Array<String>]
|
244
|
+
def consumer_tags(consumer)
|
245
|
+
messages = consumer.messages
|
246
|
+
metadata = messages.metadata
|
247
|
+
consumer_group_id = consumer.topic.consumer_group.id
|
248
|
+
|
249
|
+
[
|
250
|
+
"topic:#{metadata.topic}",
|
251
|
+
"partition:#{metadata.partition}",
|
252
|
+
"consumer_group:#{consumer_group_id}"
|
253
|
+
]
|
254
|
+
end
|
255
|
+
end
|
256
|
+
end
|
257
|
+
end
|
258
|
+
end
|
259
|
+
end
|
data/lib/karafka/version.rb
CHANGED
data.tar.gz.sig
CHANGED
Binary file
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: karafka
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 2.0.
|
4
|
+
version: 2.0.34
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Maciej Mensfeld
|
@@ -35,7 +35,7 @@ cert_chain:
|
|
35
35
|
Qf04B9ceLUaC4fPVEz10FyobjaFoY4i32xRto3XnrzeAgfEe4swLq8bQsR3w/EF3
|
36
36
|
MGU0FeSV2Yj7Xc2x/7BzLK8xQn5l7Yy75iPF+KP3vVmDHnNl
|
37
37
|
-----END CERTIFICATE-----
|
38
|
-
date: 2023-
|
38
|
+
date: 2023-03-04 00:00:00.000000000 Z
|
39
39
|
dependencies:
|
40
40
|
- !ruby/object:Gem::Dependency
|
41
41
|
name: karafka-core
|
@@ -198,6 +198,7 @@ files:
|
|
198
198
|
- lib/karafka/instrumentation/vendors/datadog/dashboard.json
|
199
199
|
- lib/karafka/instrumentation/vendors/datadog/listener.rb
|
200
200
|
- lib/karafka/instrumentation/vendors/datadog/logger_listener.rb
|
201
|
+
- lib/karafka/instrumentation/vendors/datadog/metrics_listener.rb
|
201
202
|
- lib/karafka/licenser.rb
|
202
203
|
- lib/karafka/messages/batch_metadata.rb
|
203
204
|
- lib/karafka/messages/builders/batch_metadata.rb
|
metadata.gz.sig
CHANGED
Binary file
|