Exception messages are lost in Logstash 2.4

[magnusbaeck]

Starting in Logstash 2.4 the exception messages are lost, so you're only getting a stacktrace and have to resort to reading the source code to figure out what Logstash is complaining about. Given this invalid configuration:

input { file { path => "foo/bar" } }
output { stdout { } }

Logstash 2.3.2 says this:

$ /opt/logstash/bin/logstash -f test.config
Settings: Default pipeline workers: 8
Pipeline aborted due to error {:exception=>#<ArgumentError: File paths must be absolute, relative path specified: foo/bar>, :backtrace=>["/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-input-file-2.2.5/lib/logstash/inputs/file.rb:187:in `register'", "org/jruby/RubyArray.java:1613:in `each'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-input-file-2.2.5/lib/logstash/inputs/file.rb:185:in `register'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-core-2.3.2-java/lib/logstash/pipeline.rb:330:in `start_inputs'", "org/jruby/RubyArray.java:1613:in `each'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-core-2.3.2-java/lib/logstash/pipeline.rb:329:in `start_inputs'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-core-2.3.2-java/lib/logstash/pipeline.rb:180:in `start_workers'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-core-2.3.2-java/lib/logstash/pipeline.rb:136:in `run'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-core-2.3.2-java/lib/logstash/agent.rb:465:in `start_pipeline'"], :level=>:error}
stopping pipeline {:id=>"main"}

This is a really bad error message to show to an end user who made a trivial configuration mistake, but it gets even worse in 2.4.0:

$ /opt/logstash/bin/logstash -f test.config
Settings: Default pipeline workers: 8
Pipeline aborted due to error {:exception=>"ArgumentError", :backtrace=>["/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-input-file-2.2.5/lib/logstash/inputs/file.rb:187:in `register'", "org/jruby/RubyArray.java:1613:in `each'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-input-file-2.2.5/lib/logstash/inputs/file.rb:185:in `register'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-core-2.4.0-java/lib/logstash/pipeline.rb:330:in `start_inputs'", "org/jruby/RubyArray.java:1613:in `each'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-core-2.4.0-java/lib/logstash/pipeline.rb:329:in `start_inputs'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-core-2.4.0-java/lib/logstash/pipeline.rb:180:in `start_workers'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-core-2.4.0-java/lib/logstash/pipeline.rb:136:in `run'", "/opt/logstash/vendor/bundle/jruby/1.9/gems/logstash-core-2.4.0-java/lib/logstash/agent.rb:491:in `start_pipeline'"], :level=>:error}
stopping pipeline {:id=>"main"}

[jsvd]

This change was introduced in 4e1f8258.
The message was showing since we logged the actual exception instance, which we lost when changing to e.class.name. Adding a :message => e.message will mimic the previous behaviour.

[jsvd]

closed by #6011