Bump codahale metrics

[bkrieger]

Bumping the version of codahale metrics from 3.2.5 to 4.0.2

[robert3005]

We can't do this outside of major version since this is mostly used as a library. All the other bumps I have done in this repo are only for minor versions. As you can see from circle logs it's not as trivial as just bumping the version

[bkrieger]

Ok, closing.

[schlosna]

@bkrieger curious, was there something specific you were looking to use in metrics 4.0? Agree with Rob that this would possibly cause downstream chaos, so probably want to think through consequences & tradeoffs before upgrading in libs.

From https://github.com/dropwizard/metrics/releases main changes look like:

• Compiled and targeted JDK8
• Support for running under JDK9 [SPARK-2297][UI] Make task attempt and speculation more explicit in UI. apache/spark#1236
• Move JMX reporting to the metrics-jmx module
• Add Bill of Materials for Metrics Strip '@' symbols when merging pull requests. apache/spark#1239
• Used Java 8 Time API for data formatting
• Removed unnecessary reflection hacks for HealthCheckRegistry
• Removed internal LongAdder
• Removed internal ThreadLocalRandom
• Optimized generating random numbers
• Timer.Context now implements AutoCloseable
• Upgrade Jetty integration to Jetty 9.4
• Support tracking Jersey filters in Jersey resources ([SPARK-2177][SQL] describe table result contains only one column apache/spark#1118, [SPARK-2254] [SQL] ScalaRefection should mark primitive types as non-nullable. apache/spark#1193)
• Add ResponseMetered annotation for Jersey resources (SPARK-2246: Add user-data option to EC2 scripts apache/spark#1186)
• Add a method for timing non-throwing functions. ([SPARK-2284][UI] Mark all failed tasks as failures. apache/spark#1224)
• Unnecessary clear operation for ChunkedAssociativeArray (SPARK-2186: Spark SQL DSL support for simple aggregations such as SUM and AVG apache/spark#1211)
• Add some common metric filters ([SPARK-2125] Add sort flag and move sort into shuffle implementations apache/spark#1210)
• Add possibility to subclass Timer.Context ([SPARK-2287] [SQL] Make ScalaReflection be able to handle Generic case classes. apache/spark#1226)
• Jetty9: unhandled response should be counted as 404 and not 200 (The default value of "spark.locality.wait.process" is error. apache/spark#1232)

[bkrieger]

There's nothing specific I want; I have another library that depends on 4.x and I wanted to use that library and Spark in the same application.