karafka 1.0.1 → 1.4.14
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +5 -5
- checksums.yaml.gz.sig +0 -0
- data/.coditsu/ci.yml +3 -0
- data/.console_irbrc +1 -3
- data/.diffend.yml +3 -0
- data/.github/ISSUE_TEMPLATE/bug_report.md +50 -0
- data/.github/ISSUE_TEMPLATE/feature_request.md +20 -0
- data/.github/workflows/ci.yml +76 -0
- data/.gitignore +1 -0
- data/.ruby-version +1 -1
- data/CHANGELOG.md +286 -16
- data/CODE_OF_CONDUCT.md +1 -1
- data/CONTRIBUTING.md +6 -7
- data/Gemfile +5 -2
- data/Gemfile.lock +100 -103
- data/README.md +54 -74
- data/bin/karafka +1 -1
- data/certs/mensfeld.pem +26 -0
- data/config/errors.yml +40 -5
- data/docker-compose.yml +17 -0
- data/karafka.gemspec +31 -15
- data/lib/karafka/app.rb +19 -18
- data/lib/karafka/assignment_strategies/round_robin.rb +13 -0
- data/lib/karafka/attributes_map.rb +17 -21
- data/lib/karafka/backends/inline.rb +2 -3
- data/lib/karafka/base_consumer.rb +57 -0
- data/lib/karafka/base_responder.rb +77 -31
- data/lib/karafka/cli/base.rb +4 -4
- data/lib/karafka/cli/console.rb +11 -9
- data/lib/karafka/cli/flow.rb +9 -7
- data/lib/karafka/cli/info.rb +5 -4
- data/lib/karafka/cli/install.rb +32 -8
- data/lib/karafka/cli/missingno.rb +19 -0
- data/lib/karafka/cli/server.rb +18 -16
- data/lib/karafka/cli.rb +10 -2
- data/lib/karafka/code_reloader.rb +67 -0
- data/lib/karafka/connection/{config_adapter.rb → api_adapter.rb} +71 -22
- data/lib/karafka/connection/batch_delegator.rb +55 -0
- data/lib/karafka/connection/builder.rb +23 -0
- data/lib/karafka/connection/client.rb +120 -0
- data/lib/karafka/connection/listener.rb +39 -26
- data/lib/karafka/connection/message_delegator.rb +36 -0
- data/lib/karafka/consumers/batch_metadata.rb +10 -0
- data/lib/karafka/consumers/callbacks.rb +71 -0
- data/lib/karafka/consumers/includer.rb +64 -0
- data/lib/karafka/consumers/responders.rb +24 -0
- data/lib/karafka/{controllers → consumers}/single_params.rb +3 -3
- data/lib/karafka/contracts/config.rb +21 -0
- data/lib/karafka/contracts/consumer_group.rb +211 -0
- data/lib/karafka/contracts/consumer_group_topic.rb +19 -0
- data/lib/karafka/contracts/responder_usage.rb +54 -0
- data/lib/karafka/contracts/server_cli_options.rb +31 -0
- data/lib/karafka/contracts.rb +10 -0
- data/lib/karafka/errors.rb +27 -12
- data/lib/karafka/fetcher.rb +15 -15
- data/lib/karafka/helpers/class_matcher.rb +20 -10
- data/lib/karafka/helpers/config_retriever.rb +3 -3
- data/lib/karafka/helpers/inflector.rb +26 -0
- data/lib/karafka/helpers/multi_delegator.rb +0 -1
- data/lib/karafka/instrumentation/logger.rb +54 -0
- data/lib/karafka/instrumentation/monitor.rb +70 -0
- data/lib/karafka/instrumentation/proctitle_listener.rb +36 -0
- data/lib/karafka/instrumentation/stdout_listener.rb +140 -0
- data/lib/karafka/params/batch_metadata.rb +26 -0
- data/lib/karafka/params/builders/batch_metadata.rb +30 -0
- data/lib/karafka/params/builders/params.rb +38 -0
- data/lib/karafka/params/builders/params_batch.rb +25 -0
- data/lib/karafka/params/metadata.rb +20 -0
- data/lib/karafka/params/params.rb +35 -107
- data/lib/karafka/params/params_batch.rb +38 -19
- data/lib/karafka/patches/ruby_kafka.rb +47 -0
- data/lib/karafka/persistence/client.rb +29 -0
- data/lib/karafka/persistence/consumers.rb +45 -0
- data/lib/karafka/persistence/topics.rb +48 -0
- data/lib/karafka/process.rb +6 -9
- data/lib/karafka/responders/builder.rb +15 -14
- data/lib/karafka/responders/topic.rb +14 -9
- data/lib/karafka/routing/builder.rb +38 -9
- data/lib/karafka/routing/consumer_group.rb +6 -4
- data/lib/karafka/routing/consumer_mapper.rb +10 -9
- data/lib/karafka/routing/proxy.rb +10 -1
- data/lib/karafka/routing/router.rb +1 -1
- data/lib/karafka/routing/topic.rb +8 -12
- data/lib/karafka/routing/topic_mapper.rb +16 -18
- data/lib/karafka/serialization/json/deserializer.rb +27 -0
- data/lib/karafka/serialization/json/serializer.rb +31 -0
- data/lib/karafka/server.rb +50 -39
- data/lib/karafka/setup/config.rb +138 -91
- data/lib/karafka/setup/configurators/water_drop.rb +21 -16
- data/lib/karafka/setup/dsl.rb +21 -0
- data/lib/karafka/status.rb +7 -3
- data/lib/karafka/templates/{application_controller.rb.example → application_consumer.rb.erb} +2 -2
- data/lib/karafka/templates/karafka.rb.erb +92 -0
- data/lib/karafka/version.rb +1 -1
- data/lib/karafka.rb +19 -15
- data.tar.gz.sig +0 -0
- metadata +119 -81
- metadata.gz.sig +5 -0
- data/.github/ISSUE_TEMPLATE.md +0 -2
- data/.travis.yml +0 -17
- data/Rakefile +0 -7
- data/lib/karafka/base_controller.rb +0 -117
- data/lib/karafka/connection/messages_consumer.rb +0 -106
- data/lib/karafka/connection/messages_processor.rb +0 -61
- data/lib/karafka/controllers/includer.rb +0 -51
- data/lib/karafka/controllers/responders.rb +0 -19
- data/lib/karafka/loader.rb +0 -29
- data/lib/karafka/logger.rb +0 -53
- data/lib/karafka/monitor.rb +0 -98
- data/lib/karafka/parsers/json.rb +0 -38
- data/lib/karafka/patches/dry_configurable.rb +0 -33
- data/lib/karafka/persistence/controller.rb +0 -23
- data/lib/karafka/schemas/config.rb +0 -31
- data/lib/karafka/schemas/consumer_group.rb +0 -64
- data/lib/karafka/schemas/consumer_group_topic.rb +0 -18
- data/lib/karafka/schemas/responder_usage.rb +0 -38
- data/lib/karafka/schemas/server_cli_options.rb +0 -43
- data/lib/karafka/setup/configurators/base.rb +0 -35
- data/lib/karafka/setup/configurators/celluloid.rb +0 -19
- data/lib/karafka/templates/karafka.rb.example +0 -41
- /data/lib/karafka/templates/{application_responder.rb.example → application_responder.rb.erb} +0 -0
@@ -0,0 +1,54 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Karafka
|
4
|
+
module Instrumentation
|
5
|
+
# Default logger for Event Delegator
|
6
|
+
# @note It uses ::Logger features - providing basic logging
|
7
|
+
class Logger < ::Logger
|
8
|
+
# Map containing information about log level for given environment
|
9
|
+
ENV_MAP = {
|
10
|
+
'production' => ::Logger::ERROR,
|
11
|
+
'test' => ::Logger::ERROR,
|
12
|
+
'development' => ::Logger::INFO,
|
13
|
+
'debug' => ::Logger::DEBUG,
|
14
|
+
'default' => ::Logger::INFO
|
15
|
+
}.freeze
|
16
|
+
|
17
|
+
private_constant :ENV_MAP
|
18
|
+
|
19
|
+
# Creates a new instance of logger ensuring that it has a place to write to
|
20
|
+
# @param _args Any arguments that we don't care about but that are needed in order to
|
21
|
+
# make this logger compatible with the default Ruby one
|
22
|
+
def initialize(*_args)
|
23
|
+
super(target)
|
24
|
+
self.level = ENV_MAP[Karafka.env] || ENV_MAP['default']
|
25
|
+
end
|
26
|
+
|
27
|
+
private
|
28
|
+
|
29
|
+
# @return [Karafka::Helpers::MultiDelegator] multi delegator instance
|
30
|
+
# to which we will be writing logs
|
31
|
+
# We use this approach to log stuff to file and to the $stdout at the same time
|
32
|
+
def target
|
33
|
+
Karafka::Helpers::MultiDelegator
|
34
|
+
.delegate(:write, :close)
|
35
|
+
.to(*[$stdout, file].compact)
|
36
|
+
end
|
37
|
+
|
38
|
+
# @return [Pathname] Path to a file to which we should log
|
39
|
+
def log_path
|
40
|
+
@log_path ||= Karafka::App.root.join("log/#{Karafka.env}.log")
|
41
|
+
end
|
42
|
+
|
43
|
+
# @return [File] file to which we want to write our logs
|
44
|
+
# @note File is being opened in append mode ('a')
|
45
|
+
def file
|
46
|
+
FileUtils.mkdir_p(File.dirname(log_path))
|
47
|
+
|
48
|
+
@file ||= File.open(log_path, 'a')
|
49
|
+
rescue Errno::EACCES, Errno::EROFS
|
50
|
+
nil
|
51
|
+
end
|
52
|
+
end
|
53
|
+
end
|
54
|
+
end
|
@@ -0,0 +1,70 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Karafka
|
4
|
+
# Namespace for all the things related with Karafka instrumentation process
|
5
|
+
module Instrumentation
|
6
|
+
# Monitor is used to hookup external monitoring services to monitor how Karafka works
|
7
|
+
# It provides a standardized API for checking incoming messages/enqueueing etc
|
8
|
+
# Since it is a pub-sub based on dry-monitor, you can use as many subscribers/loggers at the
|
9
|
+
# same time, which means that you might have for example file logging and NewRelic at the same
|
10
|
+
# time
|
11
|
+
# @note This class acts as a singleton because we are only permitted to have single monitor
|
12
|
+
# per running process (just as logger)
|
13
|
+
class Monitor < Dry::Monitor::Notifications
|
14
|
+
# List of events that we support in the system and to which a monitor client can hook up
|
15
|
+
# @note The non-error once support timestamp benchmarking
|
16
|
+
# @note Depending on Karafka extensions and additional engines, this might not be the
|
17
|
+
# complete list of all the events. Please use the #available_events on fully loaded
|
18
|
+
# Karafka system to determine all of the events you can use.
|
19
|
+
# Last 4 events are from WaterDrop but for convenience we use the same monitor for the
|
20
|
+
# whole karafka ecosystem
|
21
|
+
BASE_EVENTS = %w[
|
22
|
+
params.params.deserialize
|
23
|
+
params.params.deserialize.error
|
24
|
+
connection.listener.before_fetch_loop
|
25
|
+
connection.listener.fetch_loop
|
26
|
+
connection.listener.fetch_loop.error
|
27
|
+
connection.client.fetch_loop.error
|
28
|
+
connection.batch_delegator.call
|
29
|
+
connection.message_delegator.call
|
30
|
+
fetcher.call.error
|
31
|
+
backends.inline.process
|
32
|
+
process.notice_signal
|
33
|
+
consumers.responders.respond_with
|
34
|
+
async_producer.call.error
|
35
|
+
async_producer.call.retry
|
36
|
+
sync_producer.call.error
|
37
|
+
sync_producer.call.retry
|
38
|
+
app.initializing
|
39
|
+
app.initialized
|
40
|
+
app.running
|
41
|
+
app.stopping
|
42
|
+
app.stopping.error
|
43
|
+
app.stopped
|
44
|
+
].freeze
|
45
|
+
|
46
|
+
private_constant :BASE_EVENTS
|
47
|
+
|
48
|
+
# @return [Karafka::Instrumentation::Monitor] monitor instance for system instrumentation
|
49
|
+
def initialize
|
50
|
+
super(:karafka)
|
51
|
+
BASE_EVENTS.each(&method(:register_event))
|
52
|
+
end
|
53
|
+
|
54
|
+
# Allows us to subscribe to events with a code that will be yielded upon events
|
55
|
+
# @param event_name_or_listener [String, Object] name of the event we want to subscribe to
|
56
|
+
# or a listener if we decide to go with object listener
|
57
|
+
def subscribe(event_name_or_listener)
|
58
|
+
return super unless event_name_or_listener.is_a?(String)
|
59
|
+
return super if available_events.include?(event_name_or_listener)
|
60
|
+
|
61
|
+
raise Errors::UnregisteredMonitorEventError, event_name_or_listener
|
62
|
+
end
|
63
|
+
|
64
|
+
# @return [Array<String>] names of available events to which we can subscribe
|
65
|
+
def available_events
|
66
|
+
__bus__.events.keys
|
67
|
+
end
|
68
|
+
end
|
69
|
+
end
|
70
|
+
end
|
@@ -0,0 +1,36 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Karafka
|
4
|
+
module Instrumentation
|
5
|
+
# Listener that sets a proc title with a nice descriptive value
|
6
|
+
class ProctitleListener
|
7
|
+
# Updates proc title to an initializing one
|
8
|
+
# @param _event [Dry::Events::Event] event details including payload
|
9
|
+
def on_app_initializing(_event)
|
10
|
+
setproctitle('initializing')
|
11
|
+
end
|
12
|
+
|
13
|
+
# Updates proc title to a running one
|
14
|
+
# @param _event [Dry::Events::Event] event details including payload
|
15
|
+
def on_app_running(_event)
|
16
|
+
setproctitle('running')
|
17
|
+
end
|
18
|
+
|
19
|
+
# Updates proc title to a stopping one
|
20
|
+
# @param _event [Dry::Events::Event] event details including payload
|
21
|
+
def on_app_stopping(_event)
|
22
|
+
setproctitle('stopping')
|
23
|
+
end
|
24
|
+
|
25
|
+
private
|
26
|
+
|
27
|
+
# Sets a proper proc title with our constant prefix
|
28
|
+
# @param status [String] any status we want to set
|
29
|
+
def setproctitle(status)
|
30
|
+
::Process.setproctitle(
|
31
|
+
"karafka #{Karafka::App.config.client_id} (#{status})"
|
32
|
+
)
|
33
|
+
end
|
34
|
+
end
|
35
|
+
end
|
36
|
+
end
|
@@ -0,0 +1,140 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Karafka
|
4
|
+
module Instrumentation
|
5
|
+
# Default listener that hooks up to our instrumentation and uses its events for logging
|
6
|
+
# It can be removed/replaced or anything without any harm to the Karafka app flow
|
7
|
+
class StdoutListener
|
8
|
+
# Log levels that we use in this particular listener
|
9
|
+
USED_LOG_LEVELS = %i[
|
10
|
+
debug
|
11
|
+
info
|
12
|
+
error
|
13
|
+
fatal
|
14
|
+
].freeze
|
15
|
+
|
16
|
+
# Logs details about incoming batches and with which consumer we will consume them
|
17
|
+
# @param event [Dry::Events::Event] event details including payload
|
18
|
+
def on_connection_batch_delegator_call(event)
|
19
|
+
consumer = event[:consumer]
|
20
|
+
topic = consumer.topic.name
|
21
|
+
kafka_messages = event[:kafka_batch].messages
|
22
|
+
info(
|
23
|
+
<<~MSG.chomp.tr("\n", ' ')
|
24
|
+
#{kafka_messages.count} messages
|
25
|
+
on #{topic} topic
|
26
|
+
delegated to #{consumer.class}
|
27
|
+
MSG
|
28
|
+
)
|
29
|
+
end
|
30
|
+
|
31
|
+
# Logs details about incoming message and with which consumer we will consume it
|
32
|
+
# @param event [Dry::Events::Event] event details including payload
|
33
|
+
def on_connection_message_delegator_call(event)
|
34
|
+
consumer = event[:consumer]
|
35
|
+
topic = consumer.topic.name
|
36
|
+
info "1 message on #{topic} topic delegated to #{consumer.class}"
|
37
|
+
end
|
38
|
+
|
39
|
+
# Logs details about each received message value deserialization
|
40
|
+
# @param event [Dry::Events::Event] event details including payload
|
41
|
+
def on_params_params_deserialize(event)
|
42
|
+
# Keep in mind, that a caller here is a param object not a controller,
|
43
|
+
# so it returns a topic as a string, not a routing topic
|
44
|
+
debug(
|
45
|
+
<<~MSG.chomp.tr("\n", ' ')
|
46
|
+
Params deserialization for #{event[:caller].metadata.topic} topic
|
47
|
+
successful in #{event[:time]} ms
|
48
|
+
MSG
|
49
|
+
)
|
50
|
+
end
|
51
|
+
|
52
|
+
# Logs unsuccessful deserialization attempts of incoming data
|
53
|
+
# @param event [Dry::Events::Event] event details including payload
|
54
|
+
def on_params_params_deserialize_error(event)
|
55
|
+
topic = event[:caller].metadata.topic
|
56
|
+
error = event[:error]
|
57
|
+
error "Params deserialization error for #{topic} topic: #{error}"
|
58
|
+
end
|
59
|
+
|
60
|
+
# Logs errors that occurred in a listener fetch loop
|
61
|
+
# @param event [Dry::Events::Event] event details including payload
|
62
|
+
# @note It's an error as we can recover from it not a fatal
|
63
|
+
def on_connection_listener_fetch_loop_error(event)
|
64
|
+
error "Listener fetch loop error: #{event[:error]}"
|
65
|
+
end
|
66
|
+
|
67
|
+
# Logs errors that are related to the connection itself
|
68
|
+
# @param event [Dry::Events::Event] event details including payload
|
69
|
+
# @note Karafka will attempt to reconnect, so an error not a fatal
|
70
|
+
def on_connection_client_fetch_loop_error(event)
|
71
|
+
error "Client fetch loop error: #{event[:error]}"
|
72
|
+
end
|
73
|
+
|
74
|
+
# Logs info about crashed fetcher
|
75
|
+
# @param event [Dry::Events::Event] event details including payload
|
76
|
+
# @note If this happens, Karafka will shutdown as it means a critical error
|
77
|
+
# in one of the threads
|
78
|
+
def on_fetcher_call_error(event)
|
79
|
+
fatal "Fetcher crash due to an error: #{event[:error]}"
|
80
|
+
end
|
81
|
+
|
82
|
+
# Logs info about processing of a certain dataset with an inline backend
|
83
|
+
# @param event [Dry::Events::Event] event details including payload
|
84
|
+
def on_backends_inline_process(event)
|
85
|
+
count = event[:caller].send(:params_batch).to_a.size
|
86
|
+
topic = event[:caller].topic.name
|
87
|
+
time = event[:time]
|
88
|
+
info "Inline processing of topic #{topic} with #{count} messages took #{time} ms"
|
89
|
+
end
|
90
|
+
|
91
|
+
# Logs info about system signals that Karafka received
|
92
|
+
# @param event [Dry::Events::Event] event details including payload
|
93
|
+
def on_process_notice_signal(event)
|
94
|
+
info "Received #{event[:signal]} system signal"
|
95
|
+
end
|
96
|
+
|
97
|
+
# Logs info about responder usage withing a controller flow
|
98
|
+
# @param event [Dry::Events::Event] event details including payload
|
99
|
+
def on_consumers_responders_respond_with(event)
|
100
|
+
calling = event[:caller]
|
101
|
+
responder = calling.topic.responder
|
102
|
+
data = event[:data]
|
103
|
+
info "Responded from #{calling.class} using #{responder} with following data #{data}"
|
104
|
+
end
|
105
|
+
|
106
|
+
# Logs info that we're initializing Karafka framework components
|
107
|
+
# @param _event [Dry::Events::Event] event details including payload
|
108
|
+
def on_app_initializing(_event)
|
109
|
+
info "Initializing Karafka framework #{::Process.pid}"
|
110
|
+
end
|
111
|
+
|
112
|
+
# Logs info that we're running Karafka app
|
113
|
+
# @param _event [Dry::Events::Event] event details including payload
|
114
|
+
def on_app_running(_event)
|
115
|
+
info "Running Karafka server #{::Process.pid}"
|
116
|
+
end
|
117
|
+
|
118
|
+
# Logs info that we're going to stop the Karafka server
|
119
|
+
# @param _event [Dry::Events::Event] event details including payload
|
120
|
+
def on_app_stopping(_event)
|
121
|
+
# We use a separate thread as logging can't be called from trap context
|
122
|
+
Thread.new { info "Stopping Karafka server #{::Process.pid}" }
|
123
|
+
end
|
124
|
+
|
125
|
+
# Logs an error that Karafka was unable to stop the server gracefully and it had to do a
|
126
|
+
# forced exit
|
127
|
+
# @param _event [Dry::Events::Event] event details including payload
|
128
|
+
def on_app_stopping_error(_event)
|
129
|
+
# We use a separate thread as logging can't be called from trap context
|
130
|
+
Thread.new { error "Forceful Karafka server #{::Process.pid} stop" }
|
131
|
+
end
|
132
|
+
|
133
|
+
USED_LOG_LEVELS.each do |log_level|
|
134
|
+
define_method log_level do |*args|
|
135
|
+
Karafka.logger.send(log_level, *args)
|
136
|
+
end
|
137
|
+
end
|
138
|
+
end
|
139
|
+
end
|
140
|
+
end
|
@@ -0,0 +1,26 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Karafka
|
4
|
+
module Params
|
5
|
+
# Simple batch metadata object that stores all non-message information received from Kafka
|
6
|
+
# cluster while fetching the data
|
7
|
+
# @note This metadata object refers to per batch metadata, not `#params.metadata`
|
8
|
+
BatchMetadata = Struct.new(
|
9
|
+
:batch_size,
|
10
|
+
:first_offset,
|
11
|
+
:highwater_mark_offset,
|
12
|
+
:unknown_last_offset,
|
13
|
+
:last_offset,
|
14
|
+
:offset_lag,
|
15
|
+
:deserializer,
|
16
|
+
:partition,
|
17
|
+
:topic,
|
18
|
+
keyword_init: true
|
19
|
+
) do
|
20
|
+
# @return [Boolean] is the last offset known or unknown
|
21
|
+
def unknown_last_offset?
|
22
|
+
unknown_last_offset
|
23
|
+
end
|
24
|
+
end
|
25
|
+
end
|
26
|
+
end
|
@@ -0,0 +1,30 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Karafka
|
4
|
+
module Params
|
5
|
+
module Builders
|
6
|
+
# Builder for creating batch metadata object based on the batch informations
|
7
|
+
module BatchMetadata
|
8
|
+
class << self
|
9
|
+
# Creates metadata based on the kafka batch data
|
10
|
+
# @param kafka_batch [Kafka::FetchedBatch] kafka batch details
|
11
|
+
# @param topic [Karafka::Routing::Topic] topic for which we've fetched the batch
|
12
|
+
# @return [Karafka::Params::BatchMetadata] batch metadata object
|
13
|
+
def from_kafka_batch(kafka_batch, topic)
|
14
|
+
Karafka::Params::BatchMetadata.new(
|
15
|
+
batch_size: kafka_batch.messages.count,
|
16
|
+
first_offset: kafka_batch.first_offset,
|
17
|
+
highwater_mark_offset: kafka_batch.highwater_mark_offset,
|
18
|
+
unknown_last_offset: kafka_batch.unknown_last_offset?,
|
19
|
+
last_offset: kafka_batch.last_offset,
|
20
|
+
offset_lag: kafka_batch.offset_lag,
|
21
|
+
deserializer: topic.deserializer,
|
22
|
+
partition: kafka_batch.partition,
|
23
|
+
topic: topic.name
|
24
|
+
).freeze
|
25
|
+
end
|
26
|
+
end
|
27
|
+
end
|
28
|
+
end
|
29
|
+
end
|
30
|
+
end
|
@@ -0,0 +1,38 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Karafka
|
4
|
+
module Params
|
5
|
+
# Due to the fact, that we create params related objects in couple contexts / places
|
6
|
+
# plus backends can build up them their own way we have this namespace.
|
7
|
+
# It allows to isolate actual params objects from their building process that can be
|
8
|
+
# context dependent.
|
9
|
+
module Builders
|
10
|
+
# Builder for params
|
11
|
+
module Params
|
12
|
+
class << self
|
13
|
+
# @param kafka_message [Kafka::FetchedMessage] message fetched from Kafka
|
14
|
+
# @param topic [Karafka::Routing::Topic] topic for which this message was fetched
|
15
|
+
# @return [Karafka::Params::Params] params object with payload and message metadata
|
16
|
+
def from_kafka_message(kafka_message, topic)
|
17
|
+
metadata = Karafka::Params::Metadata.new(
|
18
|
+
create_time: kafka_message.create_time,
|
19
|
+
headers: kafka_message.headers || {},
|
20
|
+
is_control_record: kafka_message.is_control_record,
|
21
|
+
key: kafka_message.key,
|
22
|
+
offset: kafka_message.offset,
|
23
|
+
deserializer: topic.deserializer,
|
24
|
+
partition: kafka_message.partition,
|
25
|
+
receive_time: Time.now,
|
26
|
+
topic: topic.name
|
27
|
+
).freeze
|
28
|
+
|
29
|
+
Karafka::Params::Params.new(
|
30
|
+
kafka_message.value,
|
31
|
+
metadata
|
32
|
+
)
|
33
|
+
end
|
34
|
+
end
|
35
|
+
end
|
36
|
+
end
|
37
|
+
end
|
38
|
+
end
|
@@ -0,0 +1,25 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Karafka
|
4
|
+
module Params
|
5
|
+
module Builders
|
6
|
+
# Builder for creating params batch instances
|
7
|
+
module ParamsBatch
|
8
|
+
class << self
|
9
|
+
# Creates params batch with params inside based on the incoming messages
|
10
|
+
# and the topic from which it comes
|
11
|
+
# @param kafka_messages [Array<Kafka::FetchedMessage>] raw fetched messages
|
12
|
+
# @param topic [Karafka::Routing::Topic] topic for which we're received messages
|
13
|
+
# @return [Karafka::Params::ParamsBatch<Karafka::Params::Params>] batch with params
|
14
|
+
def from_kafka_messages(kafka_messages, topic)
|
15
|
+
params_array = kafka_messages.map do |message|
|
16
|
+
Karafka::Params::Builders::Params.from_kafka_message(message, topic)
|
17
|
+
end
|
18
|
+
|
19
|
+
Karafka::Params::ParamsBatch.new(params_array).freeze
|
20
|
+
end
|
21
|
+
end
|
22
|
+
end
|
23
|
+
end
|
24
|
+
end
|
25
|
+
end
|
@@ -0,0 +1,20 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Karafka
|
4
|
+
module Params
|
5
|
+
# Single message / params metadata details that can be accessed without the need for the
|
6
|
+
# payload deserialization
|
7
|
+
Metadata = Struct.new(
|
8
|
+
:create_time,
|
9
|
+
:headers,
|
10
|
+
:is_control_record,
|
11
|
+
:key,
|
12
|
+
:offset,
|
13
|
+
:deserializer,
|
14
|
+
:partition,
|
15
|
+
:receive_time,
|
16
|
+
:topic,
|
17
|
+
keyword_init: true
|
18
|
+
)
|
19
|
+
end
|
20
|
+
end
|
@@ -3,123 +3,51 @@
|
|
3
3
|
module Karafka
|
4
4
|
# Params namespace encapsulating all the logic that is directly related to params handling
|
5
5
|
module Params
|
6
|
-
# Class-wrapper for hash with indifferent access with additional lazy loading feature
|
7
6
|
# It provides lazy loading not only until the first usage, but also allows us to skip
|
8
|
-
# using
|
9
|
-
# heavy-
|
10
|
-
class Params
|
11
|
-
|
12
|
-
|
13
|
-
|
14
|
-
|
15
|
-
|
16
|
-
|
17
|
-
]
|
18
|
-
|
19
|
-
|
20
|
-
|
21
|
-
|
22
|
-
|
23
|
-
|
24
|
-
PARAMS_METHOD_ATTRIBUTES = %i[
|
25
|
-
topic
|
26
|
-
partition
|
27
|
-
offset
|
28
|
-
key
|
29
|
-
].freeze
|
30
|
-
|
31
|
-
class << self
|
32
|
-
# We allow building instances only via the #build method
|
33
|
-
|
34
|
-
# @param message [Kafka::FetchedMessage, Hash] message that we get out of Kafka
|
35
|
-
# in case of building params inside main Karafka process in
|
36
|
-
# Karafka::Connection::Consumer, or a hash when we retrieve data that is already parsed
|
37
|
-
# @param parser [Class] parser class that we will use to unparse data
|
38
|
-
# @return [Karafka::Params::Params] Karafka params object not yet used parser for
|
39
|
-
# retrieving data that we've got from Kafka
|
40
|
-
# @example Build params instance from a hash
|
41
|
-
# Karafka::Params::Params.build({ key: 'value' }) #=> params object
|
42
|
-
# @example Build params instance from a Kafka::FetchedMessage object
|
43
|
-
# Karafka::Params::Params.build(message) #=> params object
|
44
|
-
def build(message, parser)
|
45
|
-
# Hash case happens inside backends that interchange data
|
46
|
-
if message.is_a?(Hash)
|
47
|
-
new(parser: parser).send(:merge!, message)
|
48
|
-
else
|
49
|
-
# This happens inside Kafka::FetchedMessagesProcessor
|
50
|
-
new(
|
51
|
-
parser: parser,
|
52
|
-
parsed: false,
|
53
|
-
received_at: Time.now
|
54
|
-
).tap do |instance|
|
55
|
-
KAFKA_MESSAGE_ATTRIBUTES.each do |attribute|
|
56
|
-
instance[attribute] = message.send(attribute)
|
57
|
-
end
|
58
|
-
|
59
|
-
# When we get raw messages, they might have a topic, that was modified by a
|
60
|
-
# topic mapper. We need to "reverse" this change and map back to the non-modified
|
61
|
-
# format, so our internal flow is not corrupted with the mapping
|
62
|
-
instance[:topic] = Karafka::App.config.topic_mapper.incoming(message.topic)
|
63
|
-
end
|
64
|
-
end
|
65
|
-
end
|
66
|
-
|
67
|
-
# Defines a method call accessor to a particular hash field.
|
68
|
-
# @note Won't work for complex key names that contain spaces, etc
|
69
|
-
# @param key [Symbol] name of a field that we want to retrieve with a method call
|
70
|
-
# @example
|
71
|
-
# key_attr_reader :example
|
72
|
-
# params.example #=> 'my example value'
|
73
|
-
def key_attr_reader(key)
|
74
|
-
define_method key do
|
75
|
-
self[key]
|
76
|
-
end
|
77
|
-
end
|
7
|
+
# using deserializer until we execute our logic. That way we can operate with
|
8
|
+
# heavy-deserialization data without slowing down the whole application.
|
9
|
+
class Params
|
10
|
+
extend Forwardable
|
11
|
+
|
12
|
+
attr_reader :raw_payload, :metadata
|
13
|
+
|
14
|
+
def_delegators :metadata, *Metadata.members
|
15
|
+
|
16
|
+
# @param raw_payload [Object] incoming payload before deserialization
|
17
|
+
# @param metadata [Karafka::Params::Metadata] message metadata object
|
18
|
+
def initialize(raw_payload, metadata)
|
19
|
+
@raw_payload = raw_payload
|
20
|
+
@metadata = metadata
|
21
|
+
@deserialized = false
|
22
|
+
@payload = nil
|
78
23
|
end
|
79
24
|
|
80
|
-
# @return [
|
81
|
-
|
82
|
-
|
83
|
-
# parse it again.
|
84
|
-
def retrieve!
|
85
|
-
return self if self[:parsed]
|
25
|
+
# @return [Object] lazy-deserialized data (deserialized upon first request)
|
26
|
+
def payload
|
27
|
+
return @payload if deserialized?
|
86
28
|
|
87
|
-
|
29
|
+
@payload = deserialize
|
30
|
+
# We mark deserialization as successful after deserialization, as in case of an error
|
31
|
+
# this won't be falsely set to true
|
32
|
+
@deserialized = true
|
33
|
+
@payload
|
88
34
|
end
|
89
35
|
|
90
|
-
|
36
|
+
# @return [Boolean] did given params payload were deserialized already
|
37
|
+
def deserialized?
|
38
|
+
@deserialized
|
39
|
+
end
|
91
40
|
|
92
41
|
private
|
93
42
|
|
94
|
-
#
|
95
|
-
|
96
|
-
|
97
|
-
|
98
|
-
|
99
|
-
|
100
|
-
|
101
|
-
# @example Merge with hash with same keys (symbol based)
|
102
|
-
# new(a: 1).merge!(a: 2) #=> { a: 1 }
|
103
|
-
# @example Merge with hash with same keys (string based)
|
104
|
-
# new(a: 1).merge!('a' => 2) #=> { a: 1 }
|
105
|
-
# @example Merge with hash with same keys (current string based)
|
106
|
-
# new('a' => 1).merge!(a: 2) #=> { a: 1 }
|
107
|
-
def merge!(other_hash)
|
108
|
-
super(other_hash) { |_key, base_value, _new_value| base_value }
|
109
|
-
end
|
110
|
-
|
111
|
-
# @param value [String] Raw data that we want to parse using controller's parser
|
112
|
-
# @note If something goes wrong, it will return raw data in a hash with a message key
|
113
|
-
# @return [Hash] parsed data or a hash with message key containing raw data if something
|
114
|
-
# went wrong during parsing
|
115
|
-
def parse(value)
|
116
|
-
self[:parser].parse(value)
|
117
|
-
# We catch both of them, because for default JSON - we use JSON parser directly
|
118
|
-
rescue ::Karafka::Errors::ParserError => e
|
119
|
-
Karafka.monitor.notice_error(self.class, e)
|
43
|
+
# @return [Object] tries de-serializes data
|
44
|
+
def deserialize
|
45
|
+
Karafka.monitor.instrument('params.params.deserialize', caller: self) do
|
46
|
+
metadata.deserializer.call(self)
|
47
|
+
end
|
48
|
+
rescue ::StandardError => e
|
49
|
+
Karafka.monitor.instrument('params.params.deserialize.error', caller: self, error: e)
|
120
50
|
raise e
|
121
|
-
ensure
|
122
|
-
self[:parsed] = true
|
123
51
|
end
|
124
52
|
end
|
125
53
|
end
|