waterdrop 2.6.9 → 2.6.11

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: a4811e6ec1f3cbea82e4604e95b9cd3deddce9337e9fd935f786bd97d008e914
-  data.tar.gz: baf54ded0ca2e0ec90e659c199138a36932cd5e2f8d6afffbfdb6b770ebbd29c
+  metadata.gz: 3798d61ff3464dd6aaeb21f5457230a3b079748f260c59138142144efc62dc40
+  data.tar.gz: c9699bc202a67fbb8e456a4e0d316f56e8635e0986a09d0429a25999553c4a76
 SHA512:
-  metadata.gz: 4b2b33cfd43d78ae534c14780accb6193a6ec6ff930168c7c7e2813eb9e2599702c2db923b05793a7366f236e9d51250b07894c7cb11b8f1eca81cb79bafa3b7
-  data.tar.gz: bde20c38e6ebbac05db6085381928516c050512adffd9a4a36a14c80edfed9788b8375edc1e058f38d809cf7862d742f4116fe94efca6fddc5bb0242d4ba216f
+  metadata.gz: c6a6d03d205314775b71ca0c4e922ef52159965e30b9280727712f25368f6af78c821d6d468b2362bd4094fbd2d4747195dc1e3a7cd76a0e4582f81a66829b43
+  data.tar.gz: 9541ae05afde65cf1124661fd57e23afb4265601748c6b4a60f6fc6f8cabe463cf59b6ac683e7d047e454cc671182808f800add0f68b780bbb1b1325abcab892

data/CHANGELOG.md

@@ -1,5 +1,14 @@
 # WaterDrop changelog
 
+## 2.6.11 (2023-10-25)
+- [Enhancement] Return delivery handles and delivery report for both dummy and buffered clients with proper topics, partitions and offsets assign and auto-increment offsets per partition.
+- [Fix] Fix a case where buffered test client would not accumulate messages on failed transactions
+
+## 2.6.10 (2023-10-24)
+- [Improvement] Introduce `message.purged` event to indicate that a message that was not delivered to Kafka was purged. This most of the time refers to messages that were part of a transaction and were not yet dispatched to Kafka. It always means, that given message was not delivered but in case of transactions it is expected. In case of non-transactional it usually means `#purge` usage or exceeding `message.timeout.ms` so `librdkafka` removes this message from its internal queue. Non-transactional producers do **not** use this and pipe purges to `error.occurred`.
+- [Fix] Fix a case where `message.acknowledged` would not have `caller` key.
+- [Fix] Fix a bug where critical errors (like `IRB::Abort`) would not abort the ongoing transaction.
+
 ## 2.6.9 (2023-10-23)
 - [Improvement] Introduce a `transaction.finished` event to indicate that transaction has finished whether it was aborted or committed.
 - [Improvement] Use `transaction.committed` event to indicate that transaction has been committed.

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    waterdrop (2.6.9)
+    waterdrop (2.6.11)
       karafka-core (>= 2.2.3, < 3.0.0)
       zeitwerk (~> 2.3)
 

data/lib/waterdrop/clients/buffered.rb

@@ -6,21 +6,12 @@ module WaterDrop
     class Buffered < Clients::Dummy
       attr_accessor :messages
 
-      # Sync fake response for the message delivery to Kafka, since we do not dispatch anything
-      class SyncResponse
-        # @param _args Handler wait arguments (irrelevant as waiting is fake here)
-        def wait(*_args)
-          false
-        end
-      end
-
       # @param args [Object] anything accepted by `Clients::Dummy`
       def initialize(*args)
         super
         @messages = []
         @topics = Hash.new { |k, v| k[v] = [] }
 
-        @transaction_mutex = Mutex.new
         @transaction_active = false
         @transaction_messages = []
         @transaction_topics = Hash.new { |k, v| k[v] = [] }
@@ -29,6 +20,7 @@ module WaterDrop
 
       # "Produces" message to Kafka: it acknowledges it locally, adds it to the internal buffer
       # @param message [Hash] `WaterDrop::Producer#produce_sync` message hash
+      # @return [Dummy::Handle] fake delivery handle that can be materialized into a report
       def produce(message)
         if @transaction_active
           @transaction_topics[message.fetch(:topic)] << message
@@ -39,29 +31,20 @@ module WaterDrop
           @messages << message
         end
 
-        SyncResponse.new
+        super(**message.to_h)
       end
 
-      # Yields the code pretending it is in a transaction
-      # Supports our aborting transaction flow
-      # Moves messages the appropriate buffers only if transaction is successful
-      def transaction
+      # Starts the transaction on a given level
+      def begin_transaction
         @transaction_level += 1
-
-        return yield if @transaction_mutex.owned?
-
-        @transaction_mutex.lock
         @transaction_active = true
+      end
 
-        result = nil
-        commit = false
-
-        catch(:abort) do
-          result = yield
-          commit = true
-        end
+      # Finishes given level of transaction
+      def commit_transaction
+        @transaction_level -= 1
 
-        commit || raise(WaterDrop::Errors::AbortTransaction)
+        return unless @transaction_level.zero?
 
         # Transfer transactional data on success
         @transaction_topics.each do |topic, messages|
@@ -70,20 +53,20 @@ module WaterDrop
 
         @messages += @transaction_messages
 
-        result
-      rescue StandardError => e
-        return if e.is_a?(WaterDrop::Errors::AbortTransaction)
+        @transaction_topics.clear
+        @transaction_messages.clear
+        @transaction_active = false
+      end
 
-        raise
-      ensure
+      # Aborts the transaction
+      def abort_transaction
         @transaction_level -= 1
 
-        if @transaction_level.zero? && @transaction_mutex.owned?
-          @transaction_topics.clear
-          @transaction_messages.clear
-          @transaction_active = false
-          @transaction_mutex.unlock
-        end
+        return unless @transaction_level.zero?
+
+        @transaction_topics.clear
+        @transaction_messages.clear
+        @transaction_active = false
       end
 
       # Returns messages produced to a given topic
@@ -95,6 +78,10 @@ module WaterDrop
       # Clears internal buffer
       # Used in between specs so messages do not leak out
       def reset
+        @transaction_level = 0
+        @transaction_active = false
+        @transaction_topics.clear
+        @transaction_messages.clear
         @messages.clear
         @topics.each_value(&:clear)
       end

data/lib/waterdrop/clients/dummy.rb

@@ -5,19 +5,53 @@ module WaterDrop
     # A dummy client that is supposed to be used instead of Rdkafka::Producer in case we don't
     # want to dispatch anything to Kafka.
     #
-    # It does not store anything and just ignores messages.
+    # It does not store anything and just ignores messages. It does however return proper delivery
+    # handle that can be materialized into a report.
     class Dummy
+      # `::Rdkafka::Producer::DeliveryHandle` object API compatible dummy object
+      class Handle < ::Rdkafka::Producer::DeliveryHandle
+        # @param topic [String] topic where we want to dispatch message
+        # @param partition [Integer] target partition
+        # @param offset [Integer] offset assigned by our fake "Kafka"
+        def initialize(topic, partition, offset)
+          @topic = topic
+          @partition = partition
+          @offset = offset
+        end
+
+        # Does not wait, just creates the result
+        #
+        # @param _args [Array] anything the wait handle would accept
+        # @return [::Rdkafka::Producer::DeliveryReport]
+        def wait(*_args)
+          create_result
+        end
+
+        # Creates a delivery report with details where the message went
+        #
+        # @return [::Rdkafka::Producer::DeliveryReport]
+        def create_result
+          ::Rdkafka::Producer::DeliveryReport.new(
+            @partition,
+            @offset,
+            @topic
+          )
+        end
+      end
+
       # @param _producer [WaterDrop::Producer]
       # @return [Dummy] dummy instance
       def initialize(_producer)
-        @counter = -1
+        @counters = Hash.new { |h, k| h[k] = -1 }
       end
 
-      # Dummy method for returning the delivery report
-      # @param _args [Object] anything that the delivery handle accepts
-      # @return [::Rdkafka::Producer::DeliveryReport]
-      def wait(*_args)
-        ::Rdkafka::Producer::DeliveryReport.new(0, @counter += 1)
+      # "Produces" the message
+      # @param topic [String, Symbol] topic where we want to dispatch message
+      # @param partition [Integer] target partition
+      # @param _args [Hash] remaining details that are ignored in the dummy mode
+      # @return [Handle] delivery handle
+      def produce(topic:, partition: 0, **_args)
+        Handle.new(topic.to_s, partition, @counters["#{topic}#{partition}"] += 1)
       end
 
       # @param _args [Object] anything really, this dummy is suppose to support anything
@@ -25,26 +59,6 @@ module WaterDrop
         true
       end
 
-      # Yields the code pretending it is in a transaction
-      # Supports our aborting transaction flow
-      def transaction
-        result = nil
-        commit = false
-
-        catch(:abort) do
-          result = yield
-          commit = true
-        end
-
-        commit || raise(WaterDrop::Errors::AbortTransaction)
-
-        result
-      rescue StandardError => e
-        return if e.is_a?(WaterDrop::Errors::AbortTransaction)
-
-        raise
-      end
-
       # @param _args [Object] anything really, this dummy is suppose to support anything
       # @return [self] returns self for chaining cases
       def method_missing(*_args)

data/lib/waterdrop/clients/rdkafka.rb

@@ -19,6 +19,7 @@ module WaterDrop
           # callbacks manager to make it work
           client.delivery_callback = Instrumentation::Callbacks::Delivery.new(
             producer.id,
+            producer.transactional?,
             producer.config.monitor
           )
 

data/lib/waterdrop/instrumentation/callbacks/delivery.rb

@@ -6,19 +6,41 @@ module WaterDrop
       # Creates a callable that we want to run upon each message delivery or failure
       #
       # @note We don't have to provide client_name here as this callback is per client instance
+      #
+      # @note We do not consider `message.purge` as an error for transactional producers, because
+      #   this is a standard behaviour for not yet dispatched messages on aborted transactions.
+      #   We do however still want to instrument it for traceability.
       class Delivery
+        # Error emitted when a message was not yet dispatched and was purged from the queue
+        RD_KAFKA_RESP_PURGE_QUEUE = -152
+
+        # Error emitted when a message was purged while it was dispatched
+        RD_KAFKA_RESP_PURGE_INFLIGHT = -151
+
+        # Errors related to queue purging that is expected in transactions
+        PURGE_ERRORS = [RD_KAFKA_RESP_PURGE_INFLIGHT, RD_KAFKA_RESP_PURGE_QUEUE].freeze
+
+        private_constant :RD_KAFKA_RESP_PURGE_QUEUE, :RD_KAFKA_RESP_PURGE_INFLIGHT, :PURGE_ERRORS
+
         # @param producer_id [String] id of the current producer
+        # @param transactional [Boolean] is this handle for a transactional or regular producer
         # @param monitor [WaterDrop::Instrumentation::Monitor] monitor we are using
-        def initialize(producer_id, monitor)
+        def initialize(producer_id, transactional, monitor)
           @producer_id = producer_id
+          @transactional = transactional
           @monitor = monitor
         end
 
         # Emits delivery details to the monitor
         # @param delivery_report [Rdkafka::Producer::DeliveryReport] delivery report
         def call(delivery_report)
-          if delivery_report.error.to_i.zero?
+          error_code = delivery_report.error.to_i
+
+          if error_code.zero?
             instrument_acknowledged(delivery_report)
+
+          elsif @transactional && PURGE_ERRORS.include?(error_code)
+            instrument_purged(delivery_report)
           else
             instrument_error(delivery_report)
           end
@@ -27,24 +49,24 @@ module WaterDrop
         private
 
         # @param delivery_report [Rdkafka::Producer::DeliveryReport] delivery report
-        def instrument_error(delivery_report)
+        def instrument_acknowledged(delivery_report)
           @monitor.instrument(
-            'error.occurred',
+            'message.acknowledged',
             caller: self,
-            error: ::Rdkafka::RdkafkaError.new(delivery_report.error),
             producer_id: @producer_id,
             offset: delivery_report.offset,
             partition: delivery_report.partition,
             topic: delivery_report.topic_name,
-            delivery_report: delivery_report,
-            type: 'librdkafka.dispatch_error'
+            delivery_report: delivery_report
           )
         end
 
         # @param delivery_report [Rdkafka::Producer::DeliveryReport] delivery report
-        def instrument_acknowledged(delivery_report)
+        def instrument_purged(delivery_report)
           @monitor.instrument(
-            'message.acknowledged',
+            'message.purged',
+            caller: self,
+            error: build_error(delivery_report),
             producer_id: @producer_id,
             offset: delivery_report.offset,
             partition: delivery_report.partition,
@@ -52,6 +74,28 @@ module WaterDrop
             delivery_report: delivery_report
           )
         end
+
+        # @param delivery_report [Rdkafka::Producer::DeliveryReport] delivery report
+        def instrument_error(delivery_report)
+          @monitor.instrument(
+            'error.occurred',
+            caller: self,
+            error: build_error(delivery_report),
+            producer_id: @producer_id,
+            offset: delivery_report.offset,
+            partition: delivery_report.partition,
+            topic: delivery_report.topic_name,
+            delivery_report: delivery_report,
+            type: 'librdkafka.dispatch_error'
+          )
+        end
+
+        # Builds appropriate rdkafka error
+        # @param delivery_report [Rdkafka::Producer::DeliveryReport] delivery report
+        # @return [::Rdkafka::RdkafkaError]
+        def build_error(delivery_report)
+          ::Rdkafka::RdkafkaError.new(delivery_report.error)
+        end
       end
     end
   end

data/lib/waterdrop/instrumentation/notifications.rb

@@ -12,6 +12,7 @@ module WaterDrop
         message.produced_async
         message.produced_sync
         message.acknowledged
+        message.purged
         message.buffered
 
         messages.produced_async

data/lib/waterdrop/version.rb

@@ -3,5 +3,5 @@
 # WaterDrop library
 module WaterDrop
   # Current WaterDrop version
-  VERSION = '2.6.9'
+  VERSION = '2.6.11'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: waterdrop
 version: !ruby/object:Gem::Version
-  version: 2.6.9
+  version: 2.6.11
 platform: ruby
 authors:
 - Maciej Mensfeld
@@ -35,7 +35,7 @@ cert_chain:
   AnG1dJU+yL2BK7vaVytLTstJME5mepSZ46qqIJXMuWob/YPDmVaBF39TDSG9e34s
   msG3BiCqgOgHAnL23+CN3Rt8MsuRfEtoTKpJVcCfoEoNHOkc
   -----END CERTIFICATE-----
-date: 2023-10-23 00:00:00.000000000 Z
+date: 2023-10-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: karafka-core