Home » DevOps » Monitoring Akka with Kamon

About Nepomuk Seiler

Nepomuk Seiler is a information sience student at the Ludwigs-Maximilians-University Munich. He uses Scala for machine learning tasks and web development with play. Also he provided some small contributions to the Eclipse community.

Monitoring Akka with Kamon

I like the JVM a lot because there are a lot of tools available for inspecting a running JVM instance at runtime. The Java Mission Control (jmc) is one of my favorite tools, when it comes to monitor threads, hot methods and memory allocation.

However these tools are of limited use, when monitoring an event-driven, message-based system like Akka. A thread is almost meaningless as it could have processed any kind of message. Luckly there are some tools out there to fill this gap. Even though the Akka docs are really extensive and useful, there isn’t a lot about monitoring.

I’m more a Dev than a Ops guy, so I will only give a brief and “I thinks it does this” introduction to the monitoring-storage-gathering-displaying-stuff.

The Big Picture

First of all, when we are done we will have this infrastructure running

akka-kamon-flowcharts-single

Thanks to docker we don’t have to configure anything on the right hand-side to get started.

Kamon

Starting on the left of the picture. Kamon is a library which uses AspectJ to hook into methods calls made by the ActorSystem and record events of different types. The Kamon docs have some big gaps, but you can get a feeling of what is possible. I will not make any special configuration and just use the defaults to get started as fast as possible.

StatsD – Graphite

A network daemon that runs on the Node.js platform and listens for statistics, like counters and timers, sent over UDP and sends aggregates to one or more pluggable backend services.

Kamon provides also other backends (datadog, newrelic) to report to. For this tutorial we stick with the free StatsD server and Graphite as Backend Service.

Grafana

Grafana is a frontend for displaying your stats logged to Graphite. You have a nice Demo you can play around with. However I will give a detailed instruction on how to add your metrics in our Grafana dashboard.

Getting started

First we need an application we can monitor. I’m using my akka-kamon-activator. Checkout the code:

git clone git@github.com:muuki88/activator-akka-kamon.git

The application contains two message generators: one for peaks and one for constant load. Two types of actors handle these messages. One creates random numbers and the child actors calculate the prime factors.

Kamon Dependencies and sbt-aspectj

First we add the kamon dependencies via

val kamonVersion = "0.3.4"
 
libraryDependencies ++= Seq(
  "com.typesafe.akka" %% "akka-actor" % "2.3.5",
  "io.kamon" %% "kamon-core" % kamonVersion,
  "io.kamon" %% "kamon-statsd" % kamonVersion,
  "io.kamon" %% "kamon-log-reporter" % kamonVersion,
  "io.kamon" %% "kamon-system-metrics" % kamonVersion,
  "org.aspectj" % "aspectjweaver" % "1.8.1"
)

Next we configure the sbt-aspectj-plugin to weave our code at compile time. First add the plugin to your plugins.sbt

addSbtPlugin("com.typesafe.sbt" % "sbt-aspectj" % "0.9.4")

And now we configure it

aspectjSettings
 
javaOptions <++= AspectjKeys.weaverOptions in Aspectj
 
// when you call "sbt run" aspectj weaving kicks in
fork in run := true

Last step is to configure what should be recorded. Open up your application.conf where your akka configuration resides. Kamon uses the kamon configuration key.

kamon {
 
  # What should be recorder
  metrics {
    filters = [
      {
        # actors we should be monitored
        actor {
          includes = [ "user/*", "user/worker-*" ] # a list of what should be included
          excludes = [ "system/*" ]                # a list of what should be excluded
        }
      },
 
      # not sure about this yet. Looks important
      {
        trace {
          includes = [ "*" ]
          excludes = []
        }
      }
    ]
  }
 
  # ~~~~~~ StatsD configuration ~~~~~~~~~~~~~~~~~~~~~~~~
 
  statsd {
    # Hostname and port in which your StatsD is running. Remember that StatsD packets are sent using UDP and
    # setting unreachable hosts and/or not open ports wont be warned by the Kamon, your data wont go anywhere.
    hostname = "127.0.0.1"
    port = 8125
 
    # Interval between metrics data flushes to StatsD. It's value must be equal or greater than the
    # kamon.metrics.tick-interval setting.
    flush-interval = 1 second
 
    # Max packet size for UDP metrics data sent to StatsD.
    max-packet-size = 1024 bytes
 
    # Subscription patterns used to select which metrics will be pushed to StatsD. Note that first, metrics
    # collection for your desired entities must be activated under the kamon.metrics.filters settings.
    includes {
      actor       = [ "*" ]
      trace       = [ "*" ]
      dispatcher  = [ "*" ]
    }
 
    simple-metric-key-generator {
      # Application prefix for all metrics pushed to StatsD. The default namespacing scheme for metrics follows
      # this pattern:
      #    application.host.entity.entity-name.metric-name
      application = "yourapp"
    }
  }
}

Our app is ready to run. But first, we deploy our monitoring backend.

Monitoring Backend

As we saw in the first picture, we need a lot of stuff running in order to store our log events. The libraries and components used are most likely reasonable and you (or the more Ops than Dev guy) will have to configure it. But for the moment we just fire them up all at once in a simple docker container. I don’t put them in detached mode so I see what’s going on.

docker run -v /etc/localtime:/etc/localtime:ro -p 80:80 -p 8125:8125/udp -p 8126:8126 -p 8083:8083 -p 8086:8086 -p 8084:8084 --name kamon-grafana-dashboard muuki88/grafana_graphite:latest

My image is based on a fork from the original docker image by kamon.

Run and build the Dashboard

Now go to your running Grafana instance at localhost. You see a default, which we will use to display the average time-in-mailbox. Click on the title of the graph ( First Graph (click title to edit ). Now select the metrics like this:

akka-kamon-grafana

And that’s it!

Reference: Monitoring Akka with Kamon from our JCG partner Nepomuk Seiler at the mukis.de blog.

Do you want to know how to develop your skillset to become a Java Rockstar?

Subscribe to our newsletter to start Rocking right now!

To get you started we give you our best selling eBooks for FREE!

 

1. JPA Mini Book

2. JVM Troubleshooting Guide

3. JUnit Tutorial for Unit Testing

4. Java Annotations Tutorial

5. Java Interview Questions

6. Spring Interview Questions

7. Android UI Design

 

and many more ....

 

Receive Java & Developer job alerts in your Area

I have read and agree to the terms & conditions

 

Subscribe
Notify of
guest

This site uses Akismet to reduce spam. Learn how your comment data is processed.

0 Comments
Inline Feedbacks
View all comments