This page applies to an earlier version of the AppDynamics App IQ Platform.
For documentation on the latest version, see the 4.4 Documentation.


Skip to end of metadata
Go to start of metadata
On this page:

 

Application Analytics is built on top of the AppDynamics platform. It relies on common components of the platform and adds a few of its own. The exact components you need to install and the steps for doing so depend upon your environment and your requirements. This section outlines some of those considerations and scenarios.

SaaS or On-Premise?

The first deployment question you need to consider is whether to use SaaS or On-Premise AppDynamics Analytics. You should note the following:

  • SaaS Deployment Option is Generally Available in AppDynamics 4.0.0 Release
  • On-Premise is Generally Available in AppDynamics 4.0.1 Release. 

With SaaS, AppDynamics stores the data and hosts the server components of the system for you. In the on-premise approach, you host the components yourself, storing all data on-premises. These differences are: 

  • For SaaS, you only need to install the application agent components of the system, as described in Agent Side Components.
  • For On-Premise, you need to install the Agent Side Components and the server components described in Server Side Components. The on-premise deployment involves more setup and administration, but it allows you to retain all analytics data within your own data center.  

Agent Side Components

Whether using SaaS or On-Premise Application Analytics, you need to deploy (or enable) components that provide Analytics functionality to the parts of the system that reside in your application environment. These appear on the left side of the diagram.

As numbered, the agent components are:  

 AppDynamics App Agent: Application Analytics relies upon the same App Agent that AppDynamics APM uses. If you already use AppDynamics APM, you likely already have these deployed to your monitored applications.

 Application Analytics Agent: The Analytics Agent collects data from one or more App Agents and sends it to the Analytics Processor. It also reads and transmits log data from log files from the local machine.  For Java installations, this functionality is provided by an extension of the Standalone Machine Agent and is embedded in the Standalone Machine Agent binary distribution.  For .NET, a separate Java application called "AppDynamics Analytics Agent for .NET" must be downloaded and installed.
(info) For SaaS-based installations, the connection between the Agent and the Analytics Processor in the cloud takes place over ports 80 (HTTP) and 443 (HTTPS).  For on-premise installations, the port on which the Processor receives data from the Agent is configured during installation.

 Analytics Plugin to the App Agent: The Analytics Plugin extends App Agent functionality so that it can collect and forward data to the Analytics Agent.  It is built into the 4.0 version of the Java and .NET App Agents, but is not enabled by default.  No additional download is needed here.

(info) The Controller and the Analytics Processor are on separate hosts.

Deploying Analytics Agents to Multiple Nodes 

A real world scenario is unlikely to consist of a single monitored node, as shown in the diagram above. It usually consists of many applications deployed over many hosts. 

While the main App Agent continues to send data to the AppDynamics Controller in the normal way, the Analytics Plugin, once enabled, sends its data to the Analytics Agent component.   This component runs in a separate JVM process in the local environment or network, either as part of the Standalone Machine Agent or on its own. 

There must be at least one Analytics Agent in the monitored environment, although multiple Agent plugins collecting only transaction data can share a single Analytics Agent, as shown in the figure. However, each machine from which you wish to gather log data must have its own Standalone Machine Agent with the Analytics Agent enabled. 

In this example, hosts 1-3 are collecting only transaction data, so only App Agents are required.  Each connects to the Analytics Agent on host 4, where log information is also being collected. (The Analytics Agents could also reside on the same machines as each App Agents.) On host 5, only log data is being collected, so an Analytics Agent is also necessary there.  

(info) The Controller and the Analytics Processor are on separate hosts.

Server Side Components

To set up AppDynamics Analytics on-premise, you also need to install the server parts of the system, the AppDynamics Controller and the Analytics Processor.  

The following figure depicts the basic components of an On-Premises deployment:

 

The server components add the following to the deployment:  

 AppDynamics Controller: The heart of an AppDynamics deployment, the Controller processes and presents the information gathered by the agents. 

 Analytics Processor: The server side component of Application Analytics, it gathers and stores data from the Analytics Agents and, if you have End User Monitoring, the EUM Server, allows the Controller UI to run queries on that data.  

(info) The Controller and the Analytics Processor are on separate hosts. If you are using SaaS EUM, you must use SaaS Analytics.  If you are using the on-prem EUM Server, you must use an on-prem Analytics Processor instance.

Deployment Options Summary

To summarize the scenarios, when planning your deployment, consider the following:  

  1. Are you implementing Application Analytics as a SaaS or an on-premise solution? 
    • If SaaS, the Analytics Processor is provided as part of the SaaS service. 
    • If on premises:
      • Install the Analytics Processor on a machine different from the Controller machine.
      • Install the Analytics Processor as a cluster, supervised by one or more ZooKeeper instances (production) new in 4.0.1
  2. For each monitored machine, are you capturing log data, transaction data, or both?
    • If log data only, install either the Standalone Machine Agent (Java) or the Analytics Agent application (.NET) on the host and enable the App Agent Plugin.
    • If log data and transaction data, install either the Standalone Machine Agent (Java) or the Analytics Agent application (.NET) on the host and enable the App Agent Plugin.
    • If transaction data only, install either the Standalone Machine Agent (Java) or the Analytics Agent application (.NET) on each host or use a common Analytics Agent instance that is shared.  Enable the App Agent Plugin on each host. 

The following topics describe how to install the components: 

  • No labels