Throughput and Latency on Apache Flink

Question

I have written a very simple java program for Apache Flink and now I am interested in measuring statistics such as throughput (number of tuples processed per second) and latency (the time the program needs to process every input tuple).

 StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();

env.readTextFile("/home/LizardKing/Documents/Power/Prova.csv")
        .map(new MyMapper().writeAsCsv("/home/LizardKing/Results.csv");

JobExecutionResult res = env.execute();

I know that Flink exposes some metrics:

https://ci.apache.org/projects/flink/flink-docs-release-1.2/monitoring/metrics.html

But I am not sure how to use them in order to obtain what I want. From the link I have read that a "meter" can be used to measure the average throughput but, after having defined it, how should I use it?

What exactly are you struggling with? For throughput, you'd register a Meter in your MyMapper function, just as shown in the link you provided. You can watch the metrics live in the Flink web dashboard. — us2012
If I follow the instruction I need to implement the myMeter class, I have tried something but it does not work. If I use the DropWizard meter and try to run it in the standalone mode I have an error (java.lang.NoClassDefFoundError: com/codahale/metrics/Meter ) even if I have included the dependency in the pom.xml. — LizardKing

Sohi Sohi · Accepted Answer · 2017-06-22T11:42:27

We are running custom metrics like meter, gauge in our production streaming job running on yarn .

Here are steps :

Additional dependency to pom.xml

<dependency>
    <groupId>org.apache.flink</groupId>
    <artifactId>flink-metrics-dropwizard</artifactId>
    <version>${flink.version}</version>
</dependency>

We are using version 1.2.1

Then add meter to MyMapper class .

import org.apache.flink.api.common.JobExecutionResult;
import org.apache.flink.api.common.functions.RichMapFunction;
import org.apache.flink.configuration.Configuration;
import org.apache.flink.dropwizard.metrics.DropwizardMeterWrapper;
import org.apache.flink.metrics.Meter;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;


public class Test {


    public static void main(String[] args) throws Exception {

        StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();

        env
                .readTextFile("/home/LizardKing/Documents/Power/Prova.csv")
                .map(new MyMapper())
                .writeAsCsv("/home/LizardKing/Results.csv");

        JobExecutionResult res = env.execute();
    }


    private static class MyMapper extends RichMapFunction<String, Object> {

        private transient Meter meter;

        @Override
        public void open(Configuration parameters) throws Exception {
            super.open(parameters);
            this.meter = getRuntimeContext()
                    .getMetricGroup()
                    .meter("myMeter", new DropwizardMeterWrapper(new com.codahale.metrics.Meter()));
        }

        @Override
        public Object map(String value) throws Exception {    
            this.meter.markEvent();
            return value;
        }
    }
}

Hope this helps .

Throughput and Latency on Apache Flink

1 Answers