karafka 1.3.0 → 1.4.0.rc1

data/README.md

@@ -2,11 +2,9 @@
 
 [![Build Status](https://travis-ci.org/karafka/karafka.svg?branch=master)](https://travis-ci.org/karafka/karafka)
 
-## New release in progress!
+**Note**: Documentation presented here refers to **not yet** released Karafka `1.4.x`.
 
-**Note**: Documentation presented here refers to Karafka `1.3.0`.
-
-If you are looking for the documentation for Karafka `1.2.*`, it can be found [here](https://github.com/karafka/wiki/tree/1.2).
+If you are looking for the documentation for Karafka `1.3.x`, it can be found [here](https://github.com/karafka/wiki/tree/1.3).
 
 ## About Karafka
 
@@ -98,4 +96,4 @@ This project exists thanks to all the people who contribute.
 
 We are looking for sustainable sponsorship. If your company is relying on Karafka framework or simply want to see Karafka evolve faster to meet your requirements, please consider backing the project.
 
-Please contact [Maciej Mensfeld](mailto:maciej@coditsu.io) directly for more details.
+Please contact [Maciej Mensfeld](mailto:maciej@mensfeld.pl) directly for more details.

data/certs/mensfeld.pem

@@ -1,25 +1,25 @@
 -----BEGIN CERTIFICATE-----
 MIIEODCCAqCgAwIBAgIBATANBgkqhkiG9w0BAQsFADAjMSEwHwYDVQQDDBhtYWNp
-ZWovREM9bWVuc2ZlbGQvREM9cGwwHhcNMTkwNzMwMTQ1NDU0WhcNMjAwNzI5MTQ1
-NDU0WjAjMSEwHwYDVQQDDBhtYWNpZWovREM9bWVuc2ZlbGQvREM9cGwwggGiMA0G
-CSqGSIb3DQEBAQUAA4IBjwAwggGKAoIBgQC9fCwtaHZG2SyyNXiH8r0QbJQx/xxl
-dkvwWz9QGJO+O8rEx20FB1Ab+MVkfOscwIv5jWpmk1U9whzDPl1uFtIbgu+sk+Zb
-uQlZyK/DPN6c+/BbBL+RryTBRyvkPLoCVwm7uxc/JZ1n4AI6eF4cCZ2ieZ9QgQbU
-MQs2QPqs9hT50Ez/40GnOdadVfiDDGz+NME2C4ms0BriXwZ1tcRTfJIHe2xjIbbb
-y5qRGfsLKcgMzvLQR24olixyX1MR0s4+Wveq3QL/gBhL4veUcv+UABJA8IJR0kyB
-seHHutusiwZ1v3SjjjW1xLLrc2ARV0mgCb0WaK2T4iA3oFTGLh6Ydz8LNl31KQFv
-94nRd8IhmJxrhQ6dQ/WT9IXoa5S9lfT5lPJeINemH4/6QPABzf9W2IZlCdI9wCdB
-TBaw57MKneGAYZiKjw6OALSy2ltQUCl3RqFl3VP7n8uFy1U987Q5VIIQ3O1UUsQD
-Oe/h+r7GUU4RSPKgPlrwvW9bD/UQ+zF51v8CAwEAAaN3MHUwCQYDVR0TBAIwADAL
-BgNVHQ8EBAMCBLAwHQYDVR0OBBYEFJNIBHdfEUD7TqHqIer2YhWaWhwcMB0GA1Ud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 EQQWMBSBEm1hY2llakBtZW5zZmVsZC5wbDAdBgNVHRIEFjAUgRJtYWNpZWpAbWVu
-c2ZlbGQucGwwDQYJKoZIhvcNAQELBQADggGBAKA4eqko6BTNhlysip6rfBkVTGri
-ZXsL+kRb2hLvsQJS/kLyM21oMlu+LN0aPj3qEFR8mE/YeDD8rLAfruBRTltPNbR7
-xA5eE1gkxY5LfExUtK3b2wPqfmo7mZgfcsMwfYg/tUXw1WpBCnrhAJodpGH6SXmp
-A40qFUZst0vjiOoO+aTblIHPmMJXoZ3K42dTlNKlEiDKUWMRKSgpjjYGEYalFNWI
-hHfCz2r8L2t+dYdMZg1JGbEkq4ADGsAA8ioZIpJd7V4hI17u5TCdi7X5wh/0gN0E
-CgP+nLox3D+l2q0QuQEkayr+auFYkzTCkF+BmEk1D0Ru4mcf3F4CJvEmW4Pzbjqt
-i1tsCWPtJ4E/UUKnKaWKqGbjrjHJ0MuShYzHkodox5IOiCXIQg+1+YSzfXUV6WEK
-KJG/fhg1JV5vVDdVy6x+tv5SQ5ctU0feCsVfESi3rE3zRd+nvzE9HcZ5aXeL1UtJ
-nT5Xrioegu2w1jPyVEgyZgTZC5rvD0nNS5sFNQ==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 -----END CERTIFICATE-----

data/docker-compose.yml

@@ -0,0 +1,17 @@
+version: '2'
+services:
+  zookeeper:
+    image: wurstmeister/zookeeper
+    ports:
+      - "2181:2181"
+  kafka:
+    image: wurstmeister/kafka:1.0.1
+    ports:
+      - "9092:9092"
+    environment:
+      KAFKA_ADVERTISED_HOST_NAME: localhost
+      KAFKA_ADVERTISED_PORT: 9092
+      KAFKA_ZOOKEEPER_CONNECT: zookeeper:2181
+      KAFKA_AUTO_CREATE_TOPICS_ENABLE: 'true'
+    volumes:
+      - /var/run/docker.sock:/var/run/docker.sock

data/karafka.gemspec

@@ -11,7 +11,7 @@ Gem::Specification.new do |spec|
   spec.version     = ::Karafka::VERSION
   spec.platform    = Gem::Platform::RUBY
   spec.authors     = ['Maciej Mensfeld', 'Pavlo Vavruk', 'Adam Gwozdowski']
-  spec.email       = %w[maciej@coditsu.io pavlo.vavruk@gmail.com adam99g@gmail.com]
+  spec.email       = %w[maciej@mensfeld.pl pavlo.vavruk@gmail.com adam99g@gmail.com]
   spec.homepage    = 'https://github.com/karafka/karafka'
   spec.summary     = 'Ruby based framework for working with Apache Kafka'
   spec.description = 'Framework used to simplify Apache Kafka based Ruby applications development'
@@ -23,14 +23,13 @@ Gem::Specification.new do |spec|
   spec.add_dependency 'dry-validation', '~> 1.2'
   spec.add_dependency 'envlogic', '~> 1.1'
   spec.add_dependency 'irb', '~> 1.0'
-  spec.add_dependency 'multi_json', '>= 1.12'
   spec.add_dependency 'rake', '>= 11.3'
-  spec.add_dependency 'ruby-kafka', '>= 0.7.8'
-  spec.add_dependency 'thor', '~> 0.20'
-  spec.add_dependency 'waterdrop', '~> 1.3.0'
+  spec.add_dependency 'ruby-kafka', '>= 1.0.0'
+  spec.add_dependency 'thor', '>= 0.20'
+  spec.add_dependency 'waterdrop', '~> 1.4.0'
   spec.add_dependency 'zeitwerk', '~> 2.1'
 
-  spec.required_ruby_version = '>= 2.4.0'
+  spec.required_ruby_version = '>= 2.5.0'
 
   if $PROGRAM_NAME.end_with?('gem')
     spec.signing_key = File.expand_path('~/.ssh/gem-private_key.pem')

data/lib/karafka.rb

@@ -1,14 +1,15 @@
 # frozen_string_literal: true
 
 %w[
+  delegate
   English
   waterdrop
   kafka
   envlogic
+  json
   thor
   forwardable
   fileutils
-  multi_json
   dry-configurable
   dry-validation
   dry/events/publisher

data/lib/karafka/attributes_map.rb

@@ -52,14 +52,8 @@ module Karafka
         ignored_settings = api_adapter[:subscribe]
         defined_settings = api_adapter.values.flatten
         karafka_settings = %i[batch_fetching]
-        # This is a dirty and bad hack of dry-configurable to get keys before setting values
-        dynamically_proxied = Karafka::Setup::Config
-                              ._settings
-                              .settings
-                              .find { |s| s.name == :kafka }
-                              .value
-                              .names
-                              .to_a
+
+        dynamically_proxied = Karafka::Setup::Config.config.kafka.to_h.keys
 
         (defined_settings + dynamically_proxied).uniq + karafka_settings - ignored_settings
       end

data/lib/karafka/cli.rb

@@ -47,7 +47,7 @@ end
 if ENV['KARAFKA_CONSOLE']
   # Reloads Karafka irb console session
   def reload!
-    puts "Reloading...\n"
+    Karafka.logger.info "Reloading...\n"
     Kernel.exec Karafka::Cli::Console.command
   end
 end

data/lib/karafka/cli/flow.rb

@@ -11,19 +11,22 @@ module Karafka
       def call
         topics.each do |topic|
           any_topics = !topic.responder&.topics.nil?
+          log_messages = []
 
           if any_topics
-            puts "#{topic.name} =>"
+            log_messages << "#{topic.name} =>"
 
             topic.responder.topics.each_value do |responder_topic|
               features = []
               features << (responder_topic.required? ? 'always' : 'conditionally')
 
-              print responder_topic.name, "(#{features.join(', ')})"
+              log_messages << format(responder_topic.name, "(#{features.join(', ')})")
             end
           else
-            puts "#{topic.name} => (nothing)"
+            log_messages << "#{topic.name} => (nothing)"
           end
+
+          Karafka.logger.info(log_messages.join("\n"))
         end
       end
 
@@ -34,11 +37,11 @@ module Karafka
         Karafka::App.consumer_groups.map(&:topics).flatten.sort_by(&:name)
       end
 
-      # Prints a given value with label in a nice way
+      # Formats a given value with label in a nice way
       # @param label [String] label describing value
       # @param value [String] value that should be printed
-      def print(label, value)
-        printf "%-25s %s\n", "  - #{label}:", value
+      def format(label, value)
+        "  - #{label}:                #{value}"
       end
     end
   end

data/lib/karafka/cli/info.rb

@@ -24,7 +24,7 @@ module Karafka
           "Kafka seed brokers: #{config.kafka.seed_brokers}"
         ]
 
-        puts(info.join("\n"))
+        Karafka.logger.info(info.join("\n"))
       end
     end
   end

data/lib/karafka/cli/install.rb

@@ -13,7 +13,9 @@ module Karafka
       INSTALL_DIRS = %w[
         app/consumers
         app/responders
+        app/workers
         config
+        lib
         log
         tmp/pids
       ].freeze

data/lib/karafka/connection/api_adapter.rb

@@ -14,11 +14,12 @@ module Karafka
     module ApiAdapter
       class << self
         # Builds all the configuration settings for Kafka.new method
+        # @param consumer_group [Karafka::Routing::ConsumerGroup] consumer group details
         # @return [Array<Hash>] Array with all the client arguments including hash with all
         #   the settings required by Kafka.new method
         # @note We return array, so we can inject any arguments we want, in case of changes in the
         #   raw driver
-        def client
+        def client(consumer_group)
           # This one is a default that takes all the settings except special
           # cases defined in the map
           settings = {
@@ -26,14 +27,17 @@ module Karafka
             client_id: ::Karafka::App.config.client_id
           }
 
-          kafka_configs.each do |setting_name, setting_value|
+          kafka_configs.each_key do |setting_name|
             # All options for config adapter should be ignored as we're just interested
             # in what is left, as we want to pass all the options that are "typical"
             # and not listed in the api_adapter special cases mapping. All the values
             # from the api_adapter mapping go somewhere else, not to the client directly
             next if AttributesMap.api_adapter.values.flatten.include?(setting_name)
 
-            settings[setting_name] = setting_value
+            # Settings for each consumer group are either defined per consumer group or are
+            # inherited from the global/general settings level, thus we don't have to fetch them
+            # from the kafka settings as they are already on a consumer group level
+            settings[setting_name] = consumer_group.public_send(setting_name)
           end
 
           settings_hash = sanitize(settings)
@@ -105,11 +109,13 @@ module Karafka
           # Majority of users don't use custom topic mappers. No need to change anything when it
           # is a default mapper that does not change anything. Only some cloud providers require
           # topics to be remapped
-          return [params] if Karafka::App.config.topic_mapper.is_a?(Karafka::Routing::TopicMapper)
+          return [params.metadata] if Karafka::App.config.topic_mapper.is_a?(
+            Karafka::Routing::TopicMapper
+          )
 
           # @note We don't use tap as it is around 13% slower than non-dup version
-          dupped = params.dup
-          dupped['topic'] = Karafka::App.config.topic_mapper.outgoing(params.topic)
+          dupped = params.metadata.dup
+          dupped['topic'] = Karafka::App.config.topic_mapper.outgoing(params.metadata.topic)
           [dupped]
         end
 

data/lib/karafka/connection/batch_delegator.rb

@@ -23,7 +23,11 @@ module Karafka
           ) do
             # Due to how ruby-kafka is built, we have the metadata that is stored on the batch
             # level only available for batch consuming
-            consumer.metadata = Params::Builders::Metadata.from_kafka_batch(kafka_batch, topic)
+            consumer.batch_metadata = Params::Builders::BatchMetadata.from_kafka_batch(
+              kafka_batch,
+              topic
+            )
+
             kafka_messages = kafka_batch.messages
 
             # Depending on a case (persisted or not) we might use new consumer instance per

data/lib/karafka/connection/builder.rb

@@ -6,9 +6,11 @@ module Karafka
     module Builder
       class << self
         # Builds a Kafka::Client instance that we use to work with Kafka cluster
+        # @param consumer_group [Karafka::Routing::ConsumerGroup] consumer group for which we want
+        #   to have a new Kafka client
         # @return [::Kafka::Client] returns a Kafka client
-        def call
-          Kafka.new(*ApiAdapter.client)
+        def call(consumer_group)
+          Kafka.new(*ApiAdapter.client(consumer_group))
         end
       end
     end

data/lib/karafka/connection/client.rb

@@ -97,7 +97,7 @@ module Karafka
       def kafka_consumer
         # @note We don't cache the connection internally because we cache kafka_consumer that uses
         #   kafka client object instance
-        @kafka_consumer ||= Builder.call.consumer(
+        @kafka_consumer ||= Builder.call(consumer_group).consumer(
           *ApiAdapter.consumer(consumer_group)
         ).tap do |consumer|
           consumer_group.topics.each do |topic|

data/lib/karafka/connection/listener.rb

@@ -47,10 +47,10 @@ module Karafka
           end
         end
         # This is on purpose - see the notes for this method
-        # rubocop:disable RescueException
+        # rubocop:disable Lint/RescueException
       rescue Exception => e
         Karafka.monitor.instrument('connection.listener.fetch_loop.error', caller: self, error: e)
-        # rubocop:enable RescueException
+        # rubocop:enable Lint/RescueException
         # We can stop client without a problem, as it will reinitialize itself when running the
         # `fetch_loop` again
         @client.stop

data/lib/karafka/consumers/batch_metadata.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+module Karafka
+  module Consumers
+    # Brings the batch metadata into consumers that support batch_fetching
+    module BatchMetadata
+      attr_accessor :batch_metadata
+    end
+  end
+end

data/lib/karafka/consumers/includer.rb

@@ -16,7 +16,7 @@ module Karafka
 
           bind_backend(consumer, topic)
           bind_params(consumer, topic)
-          bind_metadata(consumer, topic)
+          bind_batch_metadata(consumer, topic)
           bind_responders(consumer, topic)
         end
 
@@ -40,13 +40,14 @@ module Karafka
           consumer.extend(SingleParams)
         end
 
-        # Adds an option to work with metadata for consumer instances that have batch fetching
+        # Adds an option to work with batch metadata for consumer instances that have
+        #   batch fetching enabled
         # @param consumer [Karafka::BaseConsumer] consumer instance
         # @param topic [Karafka::Routing::Topic] topic of a consumer class
-        def bind_metadata(consumer, topic)
+        def bind_batch_metadata(consumer, topic)
           return unless topic.batch_fetching
 
-          consumer.extend(Metadata)
+          consumer.extend(BatchMetadata)
         end
 
         # Adds responders support for topics and consumers with responders defined for them

data/lib/karafka/contracts/consumer_group.rb

@@ -178,9 +178,9 @@ module Karafka
       # @param value [String] potential RSA key value
       # @return [Boolean] is the given string a valid RSA key
       def valid_private_key?(value)
-        OpenSSL::PKey::RSA.new(value)
+        OpenSSL::PKey.read(value)
         true
-      rescue OpenSSL::PKey::RSAError
+      rescue OpenSSL::PKey::PKeyError
         false
       end
 

data/lib/karafka/contracts/server_cli_options.rb

@@ -6,6 +6,8 @@ module Karafka
     # We validate some basics + the list of consumer_groups on which we want to use, to make
     # sure that all of them are defined, plus that a pidfile does not exist
     class ServerCliOptions < Dry::Validation::Contract
+      config.messages.load_paths << File.join(Karafka.gem_root, 'config', 'errors.yml')
+
       params do
         optional(:pid).filled(:str?)
         optional(:daemon).filled(:bool?)

data/lib/karafka/helpers/class_matcher.rb

@@ -44,7 +44,7 @@ module Karafka
       # @example From Namespaced::Super2Consumer matching responder
       #   matcher.name #=> Super2Responder
       def name
-        inflected = @klass.to_s.split('::').last.to_s
+        inflected = +@klass.to_s.split('::').last.to_s
         # We inject the from into the name just in case it is missing as in a situation like
         # that it would just sanitize the name without adding the "to" postfix.
         # It could create cases when we want to build for example a responder to a consumer

data/lib/karafka/instrumentation/logger.rb

@@ -36,10 +36,11 @@ module Karafka
           .to(STDOUT, file)
       end
 
-      # Makes sure the log directory exists
+      # Makes sure the log directory exists as long as we can write to it
       def ensure_dir_exists
-        dir = File.dirname(log_path)
-        FileUtils.mkdir_p(dir) unless Dir.exist?(dir)
+        FileUtils.mkdir_p(File.dirname(log_path))
+      rescue Errno::EACCES
+        nil
       end
 
       # @return [Pathname] Path to a file to which we should log

data/lib/karafka/instrumentation/stdout_listener.rb

@@ -43,7 +43,7 @@ module Karafka
         # so it returns a topic as a string, not a routing topic
         debug(
           <<~MSG.chomp.tr("\n", ' ')
-            Params deserialization for #{event[:caller].topic} topic
+            Params deserialization for #{event[:caller].metadata.topic} topic
             successful in #{event[:time]} ms
           MSG
         )
@@ -52,7 +52,9 @@ module Karafka
       # Logs unsuccessful deserialization attempts of incoming data
       # @param event [Dry::Events::Event] event details including payload
       def on_params_params_deserialize_error(event)
-        error "Params deserialization error for #{event[:caller].topic} topic: #{event[:error]}"
+        topic = event[:caller].metadata.topic
+        error = event[:error]
+        error "Params deserialization error for #{topic} topic: #{error}"
       end
 
       # Logs errors that occurred in a listener fetch loop

data/lib/karafka/params/batch_metadata.rb

@@ -0,0 +1,26 @@
+# frozen_string_literal: true
+
+module Karafka
+  module Params
+    # Simple batch metadata object that stores all non-message information received from Kafka
+    # cluster while fetching the data
+    # @note This metadata object refers to per batch metadata, not `#params.metadata`
+    BatchMetadata = Struct.new(
+      :batch_size,
+      :first_offset,
+      :highwater_mark_offset,
+      :unknown_last_offset,
+      :last_offset,
+      :offset_lag,
+      :deserializer,
+      :partition,
+      :topic,
+      keyword_init: true
+    ) do
+      # @return [Boolean] is the last offset known or unknown
+      def unknown_last_offset?
+        unknown_last_offset
+      end
+    end
+  end
+end