ExtraHop and AppDynamics

Deployment Guide  
This guide describes how to use ExtraHop and AppDynamics to provide real-time,
per-user transaction tracing across the entire application delivery chain. ExtraHop
provides wire data analytics for IT and business operational intelligence, and
AppDynamics is the leader in agent data analytics. By integrating the two technology
data sets, enterprises can view wire data and agent data together, maintaining
consistent per-user transaction information.

This solution uses the ExtraHop Open Data Stream (ODS) capability to send real-
time events and metrics to the AppDynamics REST API. While this guide focuses on
real-time, per-user transaction tracing, there are many other ways to combine wire
data and agent analytics not covered in this guide.

Contents
Software Prerequisites .......................................................................... 2  
Network Prerequisites ........................................................................... 2  
Overview ............................................................................................ 3  
Detailed Instructions ............................................................................ 4  
1) Configure the End User Experience (EUM) headers with the
AppDynamics system ........................................................................ 4  
2) Explore and define the topology where the application lives ................ 5  
3) Implement the topology and implement the ExtraHop trigger .............. 6  
4) Edit the ExtraHop trigger to configure the specific metrics to send to
AppDynamics ................................................................................... 7  
5) Configure ExtraHop ODS for HTTP output ......................................... 8  
6) Implement and configure the AppDynamics Machine Agent ................. 9  
7) Explore the AppDynamics Event Viewer............................................ 9  
8) Explore and set up the AppDynamics Metric Browser ......................... 9  
Validating Your Installation .................................................................. 10  
Conclusion ........................................................................................ 10  
Next Steps ........................................................................................ 10  
FAQ ................................................................................................. 11  
1) Where do I run the AppDynamics Machine Agent? ........................... 11  
2) What if I cannot use the AppDynamics End User Experience header
system? ......................................................................................... 11  

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 1

Software Prerequisites
ExtraHop:
• Version 4.1 GA or later including a license for ODS for HTTP
• AppDynamics Cookie Analysis bundle from the ExtraHop documentation
site at docs.extrahop.com

AppDynamics:
• A machine agent to receive HTTP REST input
• An End User Experience license for the AppDynamics EUE system to insert
and instrument the ADRUM_BT cookie
• An installed App Dynamics Machine Agent to receive HTTP messages and
forward them to the AppDynamics Controller
• AppDynamics version 3.9 or later

Network Prerequisites
ExtraHop:
• Routing and firewall rules to allow ExtraHop devices to connect to the
AppDynamics Machine Agent HTTP listener. In our example, this is TCP
port 8293 connectivity (and is configurable on both ExtraHop and
AppDynamics to be any unique, available TCP port (not UDP)).
• A topology that can be defined through the use of some uniquely
identifiable characteristic. In order for ExtraHop to uniquely trace network
segments, the ExtraHop Application Inspection (AI) triggers must be able
to identify these segments. In our example, we define a firewall segment
by VLAN, with VLAN 1129 being the Internet side of the connections,
VLAN 1321 and 1421 being either side of the firewall, and VLAN 1121
being either side of the load balancer. These “hops” are customizable to
each environment.

Refer to AppDynamics documentation for AppDynamics requirements.

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 2

Overview
The integration has three components: the ExtraHop platform, the AppDynamics
controller, and the AppDynamics Machine Agent. This section provides an overview
of the integration before explaining the step-by-step details.

This solution uses ExtraHop triggers, user-defined JavaScript code that you can write
to extract custom metrics. In this case, triggers deliver the metrics that agents
cannot—the “opaque network” of connectivity systems such as proxies, firewalls,
routers and load balancers. Such metrics include timing deltas between hops and
network anomalies. Once the metrics are extracted, they are sent to the REST
interface of the AppDynamics Machine Agent. The Machine Agent then delivers the
information to the AppDynamics Controller.

Using the examples in this guide, you will be capturing user traffic for one complete
application environment. If you have multiple application environments, you will set
up additional triggers, agents, and proxies for each additional environment.

This guide outlines four configuration tasks:

1) Defining the topology of the application environment in the trigger, giving the
ExtraHop platform visibility into the flow of AppDynamics EUE cookies through
the network. For this task, you will be implementing the topology by
modifying the ExtraHop trigger,
2) Defining the type of metrics and events sent to AppDynamics. For this task,
you will be implementing the metrics you are interested in by modifying the
ExtraHop Trigger as well,
3) Configuring the ExtraHop Open Data Stream (ODS) for HTTP to send the
HTTP output to the AppDynamics Machine Agent. For this task you will use
the administrative interface on the ExtraHop to configure an ODS destination.
4) Configuring the AppDynamics Machine Agent to receive HTTP input. For this
task, you will be modifying the AppDynamics Machine Agent configuration file.

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 3

Detailed Instructions
As you follow these steps, have the manuals ready for reference. The instructions in
this guide are not a complete substitute for the entire product configuration. Instead,
these instructions provide the critical pieces of information to complete the
procedures.

For ExtraHop documentation, click the Help icon in the ExtraHop Web UI.

For AppDynamics documentation, click the ? (Question Mark) icon in the

AppDynamics user interface.

1) Configure the End User Experience (EUM) headers with the AppDynamics
system
This step assumes that the application being instrumented has already has the
AppDynamics application agent installed and operational. After the application agent
is installed, JavaScript and cookie injection can be done automatically using the steps
highlighted here.

Refer to the AppDynamics configuration guide for the complete information on this
configuration. The end result should be the presence of ADRUM_BT cookies within
the application stream. In general:

A) Navigate to the application you would like to instrument in the AppDynamics User
Interface.
B) Click the Configure menu item.
C) Click the Instrumentation menu item.
D) Select End User Experience from the top navigation.
E) Check the Enable End User Experience Monitoring checkbox and save the
changes.

The UI will look similar to the following:

Note: These instructions are a summary and apply to this specific version of
AppDynamics. There are also multiple ways, both manually and automatically, to
configure EUE. Refer to AppDynamics documentation for details.

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 4

2) Explore and define the topology where the application lives
After configuring your application in AppDynamics for instrumentation, the next step
is understanding the application topology in oder to implement an end-to-end
monitoring solution. In this step, we will analyze and document the flow of the
application through the network and the segments that we are interested in tracking.

The core function of the ExtraHop trigger is to record timing differences as the EUE
cookie traverses each tier. Additionally, the trigger sends network metrics to
AppDynamics. In order to accurately represent the traffic flow, we will begin by
defining the “hops”.

In our deployment, we are defining the topology using VLANs. It’s also possible to
define the topology via IP addresses or other uniquely identifiable aspects of each
segment of the devices where traffic traverses.

Refer to the following example:

The diagram above shows the traffic traversing a firewall, a proxy, a router, and a
load balancer. If the firewall VLAN is 1000, the internal side of the first proxy is VLAN
1001, and the internal side of the load balancer is VLAN 1002, then the topology
definition is as follows:

VLAN1000 :Hop1
VLAN1001: Hop2
VLAN1002: Hop3

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 5

Now, let’s look at a more specific example with the trigger code.

In the trigger snippet above, you can see that we have five hops defined: VLANs
1129, 1124, 1122, 1121, and 1123. These VLANs are defined as hops zero through
four, and they have human-readable names: internet, firewall-dmz, firewall-app,
load balancer and proxy.

In the next section, you will update the trigger with the information you gathered in
this step.

3) Implement the topology and implement the ExtraHop trigger

Now you must define the topology in the trigger and implement it. Remember that
not every hop must be defined, but you should define every hop that is practical, or
particularly problematic hops. Here you will import the trigger into the environment,
save it, and then edit the key pieces.

To configure the trigger, we will first classify the events. For this trigger, we are
interested in HTTP_REQUEST and FLOW_TICK. Second, we will define the topology
within the trigger code that we defined in step 2 above. Finally, we will assign
devices to this trigger, specifically, all of the devices in the traffic path we want to
capture.

A) Create or import the trigger and give it an appropriate name. In this example,
the name is Cookie Analysis. Included in the bundles section of the ExtraHop
documentation site (docs.extrahop.com) you will find the AppDynamics
Cookie Analysis bundle. You may simply import this bundle which will install
the trigger on your ExtraHop device.
B) Add the FLOW_TICK and HTTP_REQUEST events to the trigger.
C) Paste or edit the trigger code in the ExtraHop Trigger Editor. Edit the “Map
VLANs to hops” section with the information you gathered above, inserting
the VLAN numbers, numbering them from 0 to n, and giving each a human-
readable name. Note that the name will be the name of the metric sent to
AppDynamics.
D) Save and close the trigger.
E) Navigate to the Devices section of the ExtraHop Web UI, and search for the
devices associated with the VLANS and servers you will be tracking. Select
the checkbox next to each device and then click the Select Action drop-

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 6

 down list in the upper right corner of the window to assign the device to the
“cookie analysis” trigger.

Note: Pay particular attention to step C above. Now that you have activated the
trigger, within the trigger code, ensure that you have adjusted the section “map
VLANs to hops” to match your environment, changing the VLAN numbers, adding or
removing hops numbers, and updating the comments about which hop you are
tracking.

4) Edit the ExtraHop trigger to configure the specific metrics to send to

AppDynamics
In the trigger you created above, there are two sections, one for HTTP metrics
(captured in the HTTP_REQUEST function) and one for general TCP metrics (captured
in the FLOW_TICK function). You will now modify the user-configurable settings in
each section.

The user-configurable settings for the ODS for HTTP begin on or around line 66 with
the string: /* Start of user-configurable settings to send to AppDynamics
/*. Move your cursor to this line and edit the following variables:

A) machineAgentEventURL: Enter the URL of the Machine Agent Event

REST interface. In our case, we enter “/machineagent/event. You do
not need to adjust this variable unless AppDynamics changes their
REST API. This document uses AppDynamics version 3.9.
B) machineAgentMetricURL: Enter the URL of the Machine Agent Metric
REST interface. In our case, we enter /machineagent/metrics. You
should not need to adjust this variable unless AppDynamics changes
their REST API.
C) appName: Enter the name of the application you are sending data to
for AppDynamics. This does not have to match the name exactly, but
this is the name that will appear in the AppDynamics Metric Browser. In
our case we enter MySampleApp.
D) ExtraHop: Enter the name of your ExtraHop appliance. This name is
sent to the AppDynamics Event system and identifies the ExtraHop
appliance containing the events. In our case we enter:
myExtraHopName.
E) timestamp and clientIP: You do not need to change these but are
here as an example of other types of information you can add or
remove as you send metrics to AppDynamics.

Save your changes and repeat the same changes for the FLOW function, beginning
on or around line 165 with the string: /* Start of user configurable settings
to send to AppDynamics /*.

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 7

5) Configure ExtraHop ODS for HTTP output
The ExtraHop ODS for HTTP configuration is set in the ExtraHop Administration UI.

A) In the ExtraHop Web UI, click Settings in the navigation bar and then select
Administration.
B) Log in with the proper credentials.
C) In the Configuration, section, click Open Data Streams.
D) Click HTTP.
E) On the Open Data Stream for HTTP, click Add New in the lower-left corner.
F) Configure the following settings:
• Name: If this is the first ODS HTTP destination, this will be the default
connection. Otherwise, choose a name that will be used in your trigger. In
our example, we use the first and default connection.
• Type: We choose HTTP.
• Host: Enter the IP Address of the host where the Machine Agent is
installed. In our example we enter: 10.10.9.116
• Port: Enter the port of the AppDynamics Machine Agent. The default is
8293.
Note: Make sure your ExtraHop management interface has the ability to
reach this port, including any intermediary firewalls, routers, and host
ACLs.
• For the remaining options, use the defaults.

Your configuration will look similar to the following:

Click Save to save your configuration.

You may also click Test Settings if you have already set up the AppDynamics
Machine Agent. If you have not, return to this screen and run this at a later time.
The result of the test should be HTTP 200 OK.

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 8

6) Implement and configure the AppDynamics Machine Agent
There are three main components to configuring the AppDynamics Machine Agent.
First, install the Machine Agent must be installed. Then, configure it to activate the
HTTP listener functionality. Finally, configure the controller information.

A) Refer to the AppDynamics installation instructions to obtain and install the

Machine Agent.
B) After installation, enable the HTTP listener functionality. You must modify the
startup script to add two values: metric.http.listener=true and
metric.http.listener.port. In our installation, we add the following to
values to the Machine Agent Startup script in /etc/init.d:

AGENT_OPTIONS="$AGENT_OPTIONS -Dmetric.http.listener=true"
AGENT_OPTIONS="$AGENT_OPTIONS -Dmetric.http.listener.port=8293"

C) Edit the controller-info.xml file to associate this Machine Agent with the
application that will be receiving the metrics from ExtraHop. These are values
for our implementation:
a. controller-host: appdynamics.seadmz.example.com
b. controller-port: 8090
c. application-name: MySampleApp
d. tier-name: WebTier
e. node-name: Node1
D) Save and quit the configuration of the controller-info.xml file.
E) Start the Machine Agent and confirm that it is listening on the specific port, in
this case, TCP 8293.

7) Explore the AppDynamics Event Viewer

The AppDynamics Event View is a log-based format that shows the events sent to
the event URL of the Machine Agent.

8) Explore and set up the AppDynamics Metric Browser

The AppDynamics Metric Browser allows the real-time plotting of ExtraHop metrics
and agent data. The Metric Browser contains data sent to the metrics URL of the
Machine Agent.

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 9

Validating Your Installation
To validate the ExtraHop ODS for HTTP connectivity to the AppDynamics Machine
Agent:

1. Click Test Connectivity in the ExtaHop Open Data Stream (ODS) for HTTP
window of the ExtraHop Admin UI while the AppDynamics Machine Agent is
running. You should see an HTTP 200 OK message in the ExtraHop test response
screen. It will look like this:

2. Log in to AppDynamics and ensure that metrics and events are being transmitted
from the Machine Agent to AppDynamics. Refer to AppDynamics documentation for
more information about validating the Machine Agent.

Conclusion
The end-to-end analysis capabilities of ExtraHop and AppDynamics provide a
lightweight solution to deploy and utilize, but it does have multiple moving parts. By
combining the full view of both ExtraHop and AppDynamics data, developers,
engineers and operators can see real-time correlated wire and agent data with
visualizations provided by both vendors.

Next Steps
You have now instrumented one application environment with the combination of
ExtraHop and AppDynamics.

The next steps for this deployment could take you in a variety of directions. You may
choose to customize the dashboards on both ExtraHop and AppDynamics to provide
information relevant to your environment. Alternately, you can set up another
integration to instrument other applications with AppDynamics by repeating the
steps in this document for each environment.

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 10

FAQ
Where do I run the AppDynamics Machine Agent?
The AppDynamics Machine Proxy is a lightweight Java-based proxy that gathers
machine-level metrics for AppDynamics. You can run the AppDynamics Machine
Agent on any server that is able to run a Java virtual machine. Some of the
requirements for the Machine Agent include the following:

A) A location that can run a Java Virtual Machine

B) A location that can accept incoming TCP HTTP messages on port 8293
C) A location that can communicate with the AppDyanmics Controller on port 80
or 443

What if I cannot use the AppDynamics End User Experience header system?
Application developers may want to consider using their own unique identifiers in
place of the AppDynamics End User Experience headers. Another option may be to
use a system such as Boomerang. ExtraHop provides guidance on how to use
Boomerang for traffic tracing here: http://www.extrahop.com/solutions/by-
category/real-user-monitoring/. While metric collection with ExtraHop will be
possible, using custom EUM headers changes the nature of the solution. If it is not
possible to use EUE headers in your configuration, consult your ExtraHop or
AppDynamics representative for more information.

© 2015 ExtraHop Networks, Inc. All rights reserved. Page 11