logstash-filter-grok 4.1.1 → 4.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: a6b0f1fa69a37ff0ac0c1d44d599c1d09abb706f28b12ed8bdf0299ef82eb461
-  data.tar.gz: c6b6941aa1c16a5594ec7f30b08eaec2b5e1ddc7e2e1e41c541248520d2be852
+  metadata.gz: 9a139ac82c3147d778c2f4510cf6ab077c6c4e73adeae87a0b4058623c4a0619
+  data.tar.gz: 02300cea3c6a17e10947cf63a7dae596c1ef1fc1c5114f3834ac784536a2941a
 SHA512:
-  metadata.gz: f0917c22df8a3f0f14b684e232f5b1eb9213aa36f5450ac4e28eb3ac547bac510d3d6baf2134a77fc181fdb2f94d063b91a1cba90982f25cdceaaffe9a2b377b
-  data.tar.gz: 03c469b35434026dd05651b624122020c8f98ef4a2cf7a156a6ab03aee704575e18c3540ec5c36d6f5235609e6ee3551168e094bded7120860326b339703a97e
+  metadata.gz: e822b6f1aca31141d7c3d553167a27ea1c11ba41e39dd9553fdbecfa3472780ccca75054b2247985261634efb3263a62fb7ca87f5762e418f744ef45bf994889
+  data.tar.gz: 990df0541f9a5cdb09d6fea2060c1c68e901e7cdf1cdb0fd45cfcf32dfa0710da3eeddd7fb3400441489ccc1741f0380daf5934b15e0fbba9ccc2d4a6b370c37

data/CHANGELOG.md

@@ -1,3 +1,6 @@
+## 4.2.0
+ -  Added: support for timeout_scope [#153](https://github.com/logstash-plugins/logstash-filter-grok/pull/153)
+
 ## 4.1.1
  - Fix formatting for code sample [#148](https://github.com/logstash-plugins/logstash-filter-grok/pull/148)
 

data/docs/index.asciidoc

@@ -195,6 +195,7 @@ This plugin supports the following configuration options plus the <<plugins-{typ
 | <<plugins-{type}s-{plugin}-tag_on_failure>> |<<array,array>>|No
 | <<plugins-{type}s-{plugin}-tag_on_timeout>> |<<string,string>>|No
 | <<plugins-{type}s-{plugin}-timeout_millis>> |<<number,number>>|No
+| <<plugins-{type}s-{plugin}-timeout_scope>> |<<string,string>>|No
 |=======================================================================
 
 Also see <<plugins-{type}s-{plugin}-common-options>> for a list of options supported by all
@@ -356,6 +357,22 @@ This will never timeout early, but may take a little longer to timeout.
 Actual timeout is approximate based on a 250ms quantization.
 Set to 0 to disable timeouts
 
+[id="plugins-{type}s-{plugin}-timeout_scope"]
+===== `timeout_scope`
+
+  * Value type is <<string,string>>
+  * Default value is `"pattern"`
+  * Supported values are `"pattern"` and `"event"`
+
+When multiple patterns are provided to <<plugins-{type}s-{plugin}-match>>,
+the timeout has historically applied to _each_ pattern, incurring overhead
+for each and every pattern that is attempted; when the grok filter is
+configured with `timeout_scope => event`, the plugin instead enforces
+a single timeout across all attempted matches on the event, so it can
+achieve similar safeguard against runaway matchers with significantly
+less overhead.
+
+It's usually better to scope the timeout for the whole event.
 
 
 [id="plugins-{type}s-{plugin}-common-options"]

data/lib/logstash/filters/grok.rb

@@ -140,7 +140,7 @@
   # `SYSLOGBASE` pattern which itself is defined by other patterns.
   #
   # Another option is to define patterns _inline_ in the filter using `pattern_definitions`.
-  # This is mostly for convenience and allows user to define a pattern which can be used just in that 
+  # This is mostly for convenience and allows user to define a pattern which can be used just in that
   # filter. This newly defined patterns in `pattern_definitions` will not be available outside of that particular `grok` filter.
   #
   class LogStash::Filters::Grok < LogStash::Filters::Base
@@ -168,7 +168,7 @@
     # necessarily need to define this yourself unless you are adding additional
     # patterns. You can point to multiple pattern directories using this setting.
     # Note that Grok will read all files in the directory matching the patterns_files_glob
-    # and assume it's a pattern file (including any tilde backup files). 
+    # and assume it's a pattern file (including any tilde backup files).
     # [source,ruby]
     #     patterns_dir => ["/opt/logstash/patterns", "/opt/logstash/extra_patterns"]
     #
@@ -215,6 +215,16 @@
     # Set to 0 to disable timeouts
     config :timeout_millis, :validate => :number, :default => 30000
 
+    # When multiple patterns are provided to `match`,
+    # the timeout has historically applied to _each_ pattern, incurring overhead
+    # for each and every pattern that is attempted; when the grok filter is
+    # configured with `timeout_scope => 'event'`, the plugin instead enforces
+    # a single timeout across all attempted matches on the event, so it can
+    # achieve similar safeguard against runaway matchers with significantly
+    # less overhead.
+    # It's usually better to scope the timeout for the whole event.
+    config :timeout_scope, :validate => %w(pattern event), :default => "pattern"
+
     # Tag to apply if a grok regexp times out.
     config :tag_on_timeout, :validate => :string, :default => '_groktimeout'
 
@@ -278,10 +288,8 @@
       @match_counter = metric.counter(:matches)
       @failure_counter = metric.counter(:failures)
 
-      # divide by float to allow fractionnal seconds, the Timeout class timeout value is in seconds but the underlying
-      # executor resolution is in microseconds so fractionnal second parameter down to microseconds is possible.
-      # see https://github.com/jruby/jruby/blob/9.2.7.0/core/src/main/java/org/jruby/ext/timeout/Timeout.java#L125
-      @timeout_seconds = @timeout_millis / 1000.0
+      @timeout = @timeout_millis > 0.0 ? RubyTimeout.new(@timeout_millis) : NoopTimeout::INSTANCE
+      @matcher = ( @timeout_scope.eql?('event') ? EventTimeoutMatcher : PatternTimeoutMatcher ).new(self)
     end # def register
 
     def filter(event)
@@ -334,25 +342,56 @@
     end
 
     def match_against_groks(groks, field, input, event)
-      input = input.to_s
-      matched = false
-      groks.each do |grok|
-        # Convert anything else to string (number, hash, etc)
-        matched = grok_till_timeout(grok, field, input)
-        if matched
-          grok.capture(matched) {|field, value| handle(field, value, event)}
-          break if @break_on_match
+      # Convert anything else to string (number, hash, etc)
+      context = GrokContext.new(field, input.to_s)
+      @matcher.match(context, groks, event, @break_on_match)
+    end
+
+    # Internal (base) helper to handle the global timeout switch.
+    # @private
+    class Matcher
+
+      def initialize(filter)
+        @filter = filter
+      end
+
+      def match(context, groks, event, break_on_match)
+        matched = false
+
+        groks.each do |grok|
+          context.set_grok(grok)
+
+          matched = execute(context, grok)
+          if matched
+            grok.capture(matched) { |field, value| @filter.handle(field, value, event) }
+            break if break_on_match
+          end
         end
+
+        matched
+      end
+
+      protected
+
+      def execute(context, grok)
+        grok.execute(context.input)
       end
-      
-      matched
+
     end
 
-    def grok_till_timeout(grok, field, value)
-      begin
-        @timeout_seconds > 0.0 ? Timeout.timeout(@timeout_seconds, TimeoutError) { grok.execute(value) } : grok.execute(value)
-      rescue TimeoutError
-        raise GrokTimeoutException.new(grok, field, value)
+    # @private
+    class EventTimeoutMatcher < Matcher
+      # @override
+      def match(context, groks, event, break_on_match)
+        @filter.with_timeout(context) { super }
+      end
+    end
+
+    # @private
+    class PatternTimeoutMatcher < Matcher
+      # @override
+      def execute(context, grok)
+        @filter.with_timeout(context) { super }
       end
     end
 
@@ -378,6 +417,7 @@
         end
       end
     end
+    public :handle
 
     def patterns_files_from_paths(paths, glob)
       patternfiles = []
@@ -438,4 +478,52 @@
         end
       end
     end
+
+    def with_timeout(context, &block)
+      @timeout.exec(&block)
+    rescue TimeoutError => error
+      handle_timeout(context, error)
+    end
+    public :with_timeout
+
+    def handle_timeout(context, error)
+      raise GrokTimeoutException.new(context.grok, context.field, context.input)
+    end
+
+    # @private
+    class GrokContext
+      attr_reader :grok, :field, :input
+
+      def initialize(field, input)
+        @field = field
+        @input = input
+      end
+
+      def set_grok(grok)
+        @grok = grok
+      end
+    end
+
+    # @private
+    class NoopTimeout
+      INSTANCE = new
+
+      def exec
+        yield
+      end
+    end
+
+    # @private
+    class RubyTimeout
+      def initialize(timeout_millis)
+        # divide by float to allow fractional seconds, the Timeout class timeout value is in seconds but the underlying
+        # executor resolution is in microseconds so fractional second parameter down to microseconds is possible.
+        # see https://github.com/jruby/jruby/blob/9.2.7.0/core/src/main/java/org/jruby/ext/timeout/Timeout.java#L125
+        @timeout_seconds = timeout_millis / 1000.0
+      end
+
+      def exec(&block)
+        Timeout.timeout(@timeout_seconds, TimeoutError, &block)
+      end
+    end
   end # class LogStash::Filters::Grok

data/logstash-filter-grok.gemspec

@@ -1,7 +1,7 @@
 Gem::Specification.new do |s|
 
   s.name            = 'logstash-filter-grok'
-  s.version         = '4.1.1'
+  s.version         = '4.2.0'
   s.licenses        = ['Apache License (2.0)']
   s.summary         = "Parses unstructured event data into fields"
   s.description     = "This gem is a Logstash plugin required to be installed on top of the Logstash core pipeline using $LS_HOME/bin/logstash-plugin install gemname. This gem is not a stand-alone program"

data/spec/filters/grok_spec.rb

@@ -199,7 +199,7 @@ describe LogStash::Filters::Grok do
 
     sample "400 454.33" do
       insist { subject.get("foo") } == 400
-      insist { subject.get("foo") }.is_a?(Fixnum)
+      insist { subject.get("foo") }.is_a?(Integer)
       insist { subject.get("bar") } == 454.33
       insist { subject.get("bar") }.is_a?(Float)
     end
@@ -412,7 +412,7 @@ describe LogStash::Filters::Grok do
       filter {
         grok {
           match => {
-            message => "(.*a){30}"
+            "message" => "(.*a){30}"
           }
           timeout_millis => 100
         }
@@ -425,6 +425,51 @@ describe LogStash::Filters::Grok do
     end
   end
 
+  describe "no timeout on failure with multiple patterns (when timeout not grouped)" do
+    config <<-CONFIG
+      filter {
+        grok {
+          match => {
+            "message" => [
+              "(.*f){20}", "(.*e){20}", "(.*d){20}", "(.*c){20}", "(.*b){20}",
+              "(.*a){25}", "(.*a){24}", "(.*a){23}", "(.*a){22}", "(.*a){21}",
+              "(.*a){20}"
+            ]
+          }
+          timeout_millis => 500
+          timeout_scope => 'pattern'
+        }
+      }
+    CONFIG
+
+    sample( 'b' * 10 + 'c' * 10 + 'd' * 10 + 'e' * 10 + ' ' + 'a' * 20 ) do
+      insist { subject.get("tags") }.nil?
+    end
+  end
+
+  describe "timeout on grouped (multi-pattern) failure" do
+    config <<-CONFIG
+      filter {
+        grok {
+          match => {
+            "message" => [
+              "(.*f){20}", "(.*e){20}", "(.*d){20}", "(.*c){20}", "(.*b){20}",
+              "(.*a){25}", "(.*a){24}", "(.*a){23}", "(.*a){22}", "(.*a){21}",
+              "(.*a){20}"
+            ]
+          }
+          timeout_millis => 500
+          timeout_scope => 'event'
+        }
+      }
+    CONFIG
+
+    sample( 'b' * 10 + 'c' * 10 + 'd' * 10 + 'e' * 10 + ' ' + 'a' * 20 ) do
+      expect(subject.get("tags")).to include("_groktimeout")
+      expect(subject.get("tags")).not_to include("_grokparsefailure")
+    end
+  end
+
   describe "tagging on failure" do
     config <<-CONFIG
       filter {

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: logstash-filter-grok
 version: !ruby/object:Gem::Version
-  version: 4.1.1
+  version: 4.2.0
 platform: ruby
 authors:
 - Elastic
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2019-08-08 00:00:00.000000000 Z
+date: 2019-11-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   requirement: !ruby/object:Gem::Requirement