karafka 2.0.0.beta1 → 2.0.0.beta4

data/lib/karafka/helpers/async.rb

@@ -0,0 +1,33 @@
+# frozen_string_literal: true
+
+module Karafka
+  module Helpers
+    # Allows a given class to run async in a separate thread. Provides also few methods we may
+    # want to use to control the underlying thread
+    #
+    # @note Thread running code needs to manage it's own exceptions. If they leak out, they will
+    #   abort thread on exception.
+    module Async
+      class << self
+        # Adds forwardable to redirect thread-based control methods to the underlying thread that
+        # runs the async operations
+        #
+        # @param base [Class] class we're including this module in
+        def included(base)
+          base.extend ::Forwardable
+
+          base.def_delegators :@thread, :join, :terminate, :alive?
+        end
+      end
+
+      # Runs the `#call` method in a new thread
+      def async_call
+        @thread = Thread.new do
+          Thread.current.abort_on_exception = true
+
+          call
+        end
+      end
+    end
+  end
+end

data/lib/karafka/instrumentation/logger_listener.rb

@@ -15,16 +15,43 @@ module Karafka
 
       # Logs each messages fetching attempt
       #
-      # @param _event [Dry::Events::Event] event details including payload
-      def on_connection_listener_fetch_loop(_event)
-        info 'Receiving new messages from Kafka...'
+      # @param event [Dry::Events::Event] event details including payload
+      def on_connection_listener_fetch_loop(event)
+        listener = event[:caller]
+        info "[#{listener.id}] Polling messages..."
       end
 
       # Logs about messages that we've received from Kafka
       #
       # @param event [Dry::Events::Event] event details including payload
       def on_connection_listener_fetch_loop_received(event)
-        info "Received #{event[:messages_buffer].size} new messages from Kafka"
+        listener = event[:caller]
+        time = event[:time]
+        messages_count = event[:messages_buffer].size
+        info "[#{listener.id}] Polled #{messages_count} messages in #{time}ms"
+      end
+
+      # Prints info about the fact that a given job has started
+      #
+      # @param event [Dry::Events::Event] event details including payload
+      def on_worker_process(event)
+        job = event[:job]
+        job_type = job.class.to_s.split('::').last
+        consumer = job.executor.topic.consumer
+        topic = job.executor.topic.name
+        info "[#{job.id}] #{job_type} job for #{consumer} on #{topic} started"
+      end
+
+      # Prints info about the fact that a given job has finished
+      #
+      # @param event [Dry::Events::Event] event details including payload
+      def on_worker_processed(event)
+        job = event[:job]
+        time = event[:time]
+        job_type = job.class.to_s.split('::').last
+        consumer = job.executor.topic.consumer
+        topic = job.executor.topic.name
+        info "[#{job.id}] #{job_type} job for #{consumer} on #{topic} finished in #{time}ms"
       end
 
       # Logs info about system signals that Karafka received.
@@ -52,16 +79,14 @@ module Karafka
       #
       # @param _event [Dry::Events::Event] event details including payload
       def on_app_stopping(_event)
-        # We use a separate thread as logging can't be called from trap context
-        Thread.new { info 'Stopping Karafka server' }
+        info 'Stopping Karafka server'
       end
 
       # Logs info that we stopped the Karafka server.
       #
       # @param _event [Dry::Events::Event] event details including payload
       def on_app_stopped(_event)
-        # We use a separate thread as logging can't be called from trap context
-        Thread.new { info 'Stopped Karafka server' }
+        info 'Stopped Karafka server'
       end
 
       # There are many types of errors that can occur in many places, but we provide a single
@@ -95,8 +120,7 @@ module Karafka
           fatal "Runner crashed due to an error: #{error}"
           fatal details
         when 'app.stopping.error'
-          # We use a separate thread as logging can't be called from trap context
-          Thread.new { error 'Forceful Karafka server stop' }
+          error 'Forceful Karafka server stop'
         when 'librdkafka.error'
           error "librdkafka internal error occurred: #{error}"
           error details

data/lib/karafka/instrumentation/monitor.rb

@@ -22,7 +22,6 @@ module Karafka
         app.stopping
         app.stopped
 
-        consumer.prepared
         consumer.consumed
         consumer.revoked
         consumer.shutdown
@@ -33,6 +32,9 @@ module Karafka
         connection.listener.fetch_loop
         connection.listener.fetch_loop.received
 
+        worker.process
+        worker.processed
+
         statistics.emitted
 
         error.occurred

data/lib/karafka/licenser.rb

@@ -33,6 +33,8 @@ module Karafka
 
       return if license_config.expires_on > Date.today
 
+      raise_expired_license_token_in_dev(license_config.expires_on)
+
       notify_if_license_expired(license_config.expires_on)
     end
 
@@ -53,24 +55,41 @@ module Karafka
       )
     end
 
+    # Raises an error for test and dev environments if running pro with expired license
+    # We never want to cause any non-dev problems and we should never crash anything else than
+    # tests and development envs.
+    #
+    # @param expires_on [Date] when the license expires
+    def raise_expired_license_token_in_dev(expires_on)
+      env = Karafka::App.env
+
+      return unless env.development? || env.test?
+
+      raise Errors::ExpiredLicenseTokenError.new, expired_message(expires_on)
+    end
+
     # We do not raise an error here as we don't want to cause any problems to someone that runs
     # Karafka on production. Error message is enough.
     #
     # @param expires_on [Date] when the license expires
     def notify_if_license_expired(expires_on)
-      message = <<~MSG.tr("\n", ' ')
-        Your license expired on #{expires_on}.
-        Please reach us at contact@karafka.io or visit https://karafka.io to obtain a valid one.
-      MSG
-
-      Karafka.logger.error(message)
+      Karafka.logger.error(expired_message(expires_on))
 
       Karafka.monitor.instrument(
         'error.occurred',
         caller: self,
-        error: Errors::ExpiredLicenseTokenError.new(message),
+        error: Errors::ExpiredLicenseTokenError.new(expired_message(expires_on)),
         type: 'licenser.expired'
       )
     end
+
+    # @param expires_on [Date] when the license expires
+    # @return [String] expired message
+    def expired_message(expires_on)
+      <<~MSG.tr("\n", ' ')
+        Your license expired on #{expires_on}.
+        Please reach us at contact@karafka.io or visit https://karafka.io to obtain a valid one.
+      MSG
+    end
   end
 end

data/lib/karafka/messages/batch_metadata.rb

@@ -13,10 +13,33 @@ module Karafka
       :deserializer,
       :partition,
       :topic,
+      :created_at,
       :scheduled_at,
-      :consumption_lag,
-      :processing_lag,
+      :processed_at,
       keyword_init: true
-    )
+    ) do
+      # This lag describes how long did it take for a message to be consumed from the moment it was
+      # created
+      def consumption_lag
+        time_distance_in_ms(processed_at, created_at)
+      end
+
+      # This lag describes how long did a batch have to wait before it was picked up by one of the
+      # workers
+      def processing_lag
+        time_distance_in_ms(processed_at, scheduled_at)
+      end
+
+      private
+
+      # Computes time distance in between two times in ms
+      #
+      # @param time1 [Time]
+      # @param time2 [Time]
+      # @return [Integer] distance in between two times in ms
+      def time_distance_in_ms(time1, time2)
+        ((time1 - time2) * 1_000).round
+      end
+    end
   end
 end

data/lib/karafka/messages/builders/batch_metadata.rb

@@ -8,42 +8,30 @@ module Karafka
         class << self
           # Creates metadata based on the kafka batch data.
           #
-          # @param kafka_batch [Array<Rdkafka::Consumer::Message>] raw fetched messages
+          # @param messages [Array<Karafka::Messages::Message>] messages array
           # @param topic [Karafka::Routing::Topic] topic for which we've fetched the batch
           # @param scheduled_at [Time] moment when the batch was scheduled for processing
           # @return [Karafka::Messages::BatchMetadata] batch metadata object
-          # @note Regarding the time lags: we can use the current time here, as batch metadata is
-          #   created in the worker. So whenever this is being built, it means that the processing
-          #   of this batch has already started.
-          def call(kafka_batch, topic, scheduled_at)
-            now = Time.now
-
+          #
+          # @note We do not set `processed_at` as this needs to be assigned when the batch is
+          #   picked up for processing.
+          def call(messages, topic, scheduled_at)
             Karafka::Messages::BatchMetadata.new(
-              size: kafka_batch.count,
-              first_offset: kafka_batch.first.offset,
-              last_offset: kafka_batch.last.offset,
+              size: messages.count,
+              first_offset: messages.first.offset,
+              last_offset: messages.last.offset,
               deserializer: topic.deserializer,
-              partition: kafka_batch[0].partition,
+              partition: messages.first.partition,
               topic: topic.name,
+              # We go with the assumption that the creation of the whole batch is the last message
+              # creation time
+              created_at: messages.last.timestamp,
+              # When this batch was built and scheduled for execution
               scheduled_at: scheduled_at,
-              # This lag describes how long did it take for a message to be consumed from the
-              # moment it was created
-              consumption_lag: time_distance_in_ms(now, kafka_batch.last.timestamp),
-              # This lag describes how long did a batch have to wait before it was picked up by
-              # one of the workers
-              processing_lag: time_distance_in_ms(now, scheduled_at)
-            ).freeze
-          end
-
-          private
-
-          # Computes time distance in between two times in ms
-          #
-          # @param time1 [Time]
-          # @param time2 [Time]
-          # @return [Integer] distance in between two times in ms
-          def time_distance_in_ms(time1, time2)
-            ((time1 - time2) * 1_000).round
+              # We build the batch metadata when we pick up the job in the worker, thus we can use
+              # current time here
+              processed_at: Time.now
+            )
           end
         end
       end

data/lib/karafka/messages/builders/message.rb

@@ -26,6 +26,7 @@ module Karafka
               received_at: received_at
             ).freeze
 
+            # Karafka messages cannot be frozen because of the lazy deserialization feature
             Karafka::Messages::Message.new(
               kafka_message.payload,
               metadata

data/lib/karafka/messages/builders/messages.rb

@@ -9,27 +9,19 @@ module Karafka
           # Creates messages batch with messages inside based on the incoming messages and the
           # topic from which it comes.
           #
-          # @param kafka_messages [Array<Rdkafka::Consumer::Message>] raw fetched messages
+          # @param messages [Array<Karafka::Messages::Message>] karafka messages array
           # @param topic [Karafka::Routing::Topic] topic for which we're received messages
           # @param received_at [Time] moment in time when the messages were received
           # @return [Karafka::Messages::Messages] messages batch object
-          def call(kafka_messages, topic, received_at)
-            messages_array = kafka_messages.map do |message|
-              Karafka::Messages::Builders::Message.call(
-                message,
-                topic,
-                received_at
-              )
-            end
-
+          def call(messages, topic, received_at)
             metadata = BatchMetadata.call(
-              kafka_messages,
+              messages,
               topic,
               received_at
             ).freeze
 
             Karafka::Messages::Messages.new(
-              messages_array,
+              messages,
               metadata
             ).freeze
           end

data/lib/karafka/pro/active_job/consumer.rb

@@ -0,0 +1,49 @@
+# frozen_string_literal: true
+
+# This Karafka component is a Pro component.
+# All of the commercial components are present in the lib/karafka/pro directory of this
+# repository and their usage requires commercial license agreement.
+#
+# Karafka has also commercial-friendly license, commercial support and commercial components.
+#
+# By sending a pull request to the pro components, you are agreeing to transfer the copyright of
+# your code to Maciej Mensfeld.
+
+module Karafka
+  module Pro
+    module ActiveJob
+      # Pro ActiveJob consumer that is suppose to handle long-running jobs as well as short
+      # running jobs
+      #
+      # When in LRJ, it will pause a given partition forever and will resume its processing only
+      # when all the jobs are done processing.
+      #
+      # It contains slightly better revocation warranties than the regular blocking consumer as
+      # it can stop processing batch of jobs in the middle after the revocation.
+      class Consumer < Karafka::Pro::BaseConsumer
+        # Runs ActiveJob jobs processing and handles lrj if needed
+        def consume
+          messages.each do |message|
+            # If for any reason we've lost this partition, not worth iterating over new messages
+            # as they are no longer ours
+            return if revoked?
+            break if Karafka::App.stopping?
+
+            ::ActiveJob::Base.execute(
+              ::ActiveSupport::JSON.decode(message.raw_payload)
+            )
+
+            mark_as_consumed(message)
+
+            # We check it twice as the job may be long running
+            # If marking fails, it also means it got revoked and we can stop consuming
+            return if revoked?
+
+            # Do not process more if we are shutting down
+            break if Karafka::App.stopping?
+          end
+        end
+      end
+    end
+  end
+end

data/lib/karafka/pro/active_job/dispatcher.rb

@@ -1,24 +1,24 @@
 # frozen_string_literal: true
 
+# This Karafka component is a Pro component.
+# All of the commercial components are present in the lib/karafka/pro directory of this
+# repository and their usage requires commercial license agreement.
+#
+# Karafka has also commercial-friendly license, commercial support and commercial components.
+#
+# By sending a pull request to the pro components, you are agreeing to transfer the copyright of
+# your code to Maciej Mensfeld.
+
 module Karafka
   module Pro
     # Karafka Pro ActiveJob components
     module ActiveJob
-      # This Karafka component is a Pro component.
-      # All of the commercial components are present in the lib/karafka/pro directory of this
-      # repository and their usage requires commercial license agreement.
-      #
-      # Karafka has also commercial-friendly license, commercial support and commercial components.
-      #
-      # By sending a pull request to the pro components, you are agreeing to transfer the copyright
-      # of your code to Maciej Mensfeld.
-
       # Pro dispatcher that sends the ActiveJob job to a proper topic based on the queue name
       # and that allows to inject additional options into the producer, effectively allowing for a
       # much better and more granular control over the dispatch and consumption process.
       class Dispatcher < ::Karafka::ActiveJob::Dispatcher
         # Defaults for dispatching
-        # The can be updated by using `#karafka_options` on the job
+        # They can be updated by using `#karafka_options` on the job
         DEFAULTS = {
           dispatch_method: :produce_async,
           # We don't create a dummy proc based partitioner as we would have to evaluate it with

data/lib/karafka/pro/active_job/job_options_contract.rb

@@ -1,17 +1,17 @@
 # frozen_string_literal: true
 
+# This Karafka component is a Pro component.
+# All of the commercial components are present in the lib/karafka/pro directory of this
+# repository and their usage requires commercial license agreement.
+#
+# Karafka has also commercial-friendly license, commercial support and commercial components.
+#
+# By sending a pull request to the pro components, you are agreeing to transfer the copyright of
+# your code to Maciej Mensfeld.
+
 module Karafka
   module Pro
     module ActiveJob
-      # This Karafka component is a Pro component.
-      # All of the commercial components are present in the lib/karafka/pro directory of this
-      # repository and their usage requires commercial license agreement.
-      #
-      # Karafka has also commercial-friendly license, commercial support and commercial components.
-      #
-      # By sending a pull request to the pro components, you are agreeing to transfer the copyright
-      # of your code to Maciej Mensfeld.
-
       # Contract for validating the options that can be altered with `#karafka_options` per job
       # class that works with Pro features.
       class JobOptionsContract < ::Karafka::ActiveJob::JobOptionsContract

data/lib/karafka/pro/base_consumer.rb

@@ -0,0 +1,76 @@
+# frozen_string_literal: true
+
+# This Karafka component is a Pro component.
+# All of the commercial components are present in the lib/karafka/pro directory of this
+# repository and their usage requires commercial license agreement.
+#
+# Karafka has also commercial-friendly license, commercial support and commercial components.
+#
+# By sending a pull request to the pro components, you are agreeing to transfer the copyright of
+# your code to Maciej Mensfeld.
+
+module Karafka
+  module Pro
+    # Karafka PRO consumer.
+    #
+    # If you use PRO, all your consumers should inherit (indirectly) from it.
+    #
+    # @note In case of using lrj, manual pausing may not be the best idea as resume needs to happen
+    #   after each batch is processed.
+    class BaseConsumer < Karafka::BaseConsumer
+      # Pause for tops 31 years
+      MAX_PAUSE_TIME = 1_000_000_000_000
+
+      private_constant :MAX_PAUSE_TIME
+
+      # Pauses processing of a given partition until we're done with the processing
+      # This ensures, that we can easily poll not reaching the `max.poll.interval`
+      def on_before_consume
+        # Pause at the first message in a batch. That way in case of a crash, we will not loose
+        # any messages
+        return unless topic.long_running_job?
+
+        pause(messages.first.offset, MAX_PAUSE_TIME)
+      end
+
+      # Runs extra logic after consumption that is related to handling long running jobs
+      # @note This overwrites the '#on_after_consume' from the base consumer
+      def on_after_consume
+        # Nothing to do if we lost the partition
+        return if revoked?
+
+        if @consumption.success?
+          pause_tracker.reset
+
+          # We use the non-blocking one here. If someone needs the blocking one, can implement it
+          # with manual offset management
+          # Mark as consumed only if manual offset management is not on
+          mark_as_consumed(messages.last) unless topic.manual_offset_management?
+
+          # If this is not a long running job there is nothing for us to do here
+          return unless topic.long_running_job?
+
+          # Once processing is done, we move to the new offset based on commits
+          # Here, in case manual offset management is off, we have the new proper offset of a
+          # first message from another batch from `@seek_offset`. If manual offset management
+          # is on, we move to place where the user indicated it was finished.
+          seek(@seek_offset || messages.first.offset)
+          resume
+        else
+          # If processing failed, we need to pause
+          pause(@seek_offset || messages.first.offset)
+        end
+      end
+
+      # Marks this consumer revoked state as true
+      # This allows us for things like lrj to finish early as this state may change during lrj
+      # execution
+      def on_revoked
+        # @note This may already be set to true if we tried to commit offsets and failed. In case
+        # like this it will automatically be marked as revoked.
+        @revoked = true
+        super
+      end
+    end
+  end
+end

data/lib/karafka/pro/loader.rb

@@ -1,31 +1,48 @@
 # frozen_string_literal: true
 
+# This Karafka component is a Pro component.
+# All of the commercial components are present in the lib/karafka/pro directory of this
+# repository and their usage requires commercial license agreement.
+#
+# Karafka has also commercial-friendly license, commercial support and commercial components.
+#
+# By sending a pull request to the pro components, you are agreeing to transfer the copyright of
+# your code to Maciej Mensfeld.
+
 module Karafka
   module Pro
-    # This Karafka component is a Pro component.
-    # All of the commercial components are present in the lib/karafka/pro directory of this
-    # repository and their usage requires commercial license agreement.
-    #
-    # Karafka has also commercial-friendly license, commercial support and commercial components.
-    #
-    # By sending a pull request to the pro components, you are agreeing to transfer the copyright
-    # of your code to Maciej Mensfeld.
-
     # Loader requires and loads all the pro components only when they are needed
     class Loader
+      # All the pro components that need to be loaded
+      COMPONENTS = %w[
+        base_consumer
+        performance_tracker
+        scheduler
+        processing/jobs/consume_non_blocking
+        processing/jobs_builder
+        routing/extensions
+        active_job/consumer
+        active_job/dispatcher
+        active_job/job_options_contract
+      ].freeze
+
+      private_constant :COMPONENTS
+
       class << self
         # Loads all the pro components and configures them wherever it is expected
         # @param config [Dry::Configurable::Config] whole app config that we can alter with pro
         #   components
         def setup(config)
-          require_relative 'performance_tracker'
-          require_relative 'active_job/dispatcher'
-          require_relative 'active_job/job_options_contract'
+          COMPONENTS.each { |component| require_relative(component) }
 
+          config.internal.scheduler = Scheduler.new
+          config.internal.jobs_builder = Processing::JobsBuilder.new
+          config.internal.active_job.consumer = ActiveJob::Consumer
           config.internal.active_job.dispatcher = ActiveJob::Dispatcher.new
           config.internal.active_job.job_options_contract = ActiveJob::JobOptionsContract.new
 
-          # Monitor time needed to process each message from a single partition
+          ::Karafka::Routing::Topic.include(Routing::Extensions)
+
           config.monitor.subscribe(PerformanceTracker.instance)
         end
       end

data/lib/karafka/pro/performance_tracker.rb

@@ -1,16 +1,16 @@
 # frozen_string_literal: true
 
+# This Karafka component is a Pro component.
+# All of the commercial components are present in the lib/karafka/pro directory of this
+# repository and their usage requires commercial license agreement.
+#
+# Karafka has also commercial-friendly license, commercial support and commercial components.
+#
+# By sending a pull request to the pro components, you are agreeing to transfer the copyright of
+# your code to Maciej Mensfeld.
+
 module Karafka
   module Pro
-    # This Karafka component is a Pro component.
-    # All of the commercial components are present in the lib/karafka/pro directory of this
-    # repository and their usage requires commercial license agreement.
-    #
-    # Karafka has also commercial-friendly license, commercial support and commercial components.
-    #
-    # By sending a pull request to the pro components, you are agreeing to transfer the copyright
-    # of your code to Maciej Mensfeld.
-
     # Tracker used to keep track of performance metrics
     # It provides insights that can be used to optimize processing flow
     class PerformanceTracker

data/lib/karafka/pro/processing/jobs/consume_non_blocking.rb

@@ -0,0 +1,37 @@
+# frozen_string_literal: true
+
+# This Karafka component is a Pro component.
+# All of the commercial components are present in the lib/karafka/pro directory of this
+# repository and their usage requires commercial license agreement.
+#
+# Karafka has also commercial-friendly license, commercial support and commercial components.
+#
+# By sending a pull request to the pro components, you are agreeing to transfer the copyright of
+# your code to Maciej Mensfeld.
+
+module Karafka
+  module Pro
+    # Pro components related to processing part of Karafka
+    module Processing
+      # Pro jobs
+      module Jobs
+        # The main job type in a non-blocking variant.
+        # This variant works "like" the regular consumption but pauses the partition for as long
+        # as it is needed until a job is done.
+        #
+        # It can be useful when having long lasting jobs that would exceed `max.poll.interval`
+        # if would block.
+        #
+        # @note It needs to be working with a proper consumer that will handle the partition
+        #   management. This layer of the framework knows nothing about Kafka messages consumption.
+        class ConsumeNonBlocking < ::Karafka::Processing::Jobs::Consume
+          # Releases the blocking lock after it is done with the preparation phase for this job
+          def before_call
+            super
+            @non_blocking = true
+          end
+        end
+      end
+    end
+  end
+end