A Java monitoring extension enables the AppDynamics Machine Agent to collect custom metrics, which you define and provide, and to report them to the Controller. This is an alternative to adding monitoring extensions using scripts.
When you capture custom metrics with a monitoring extension, they are supported by the same AppDynamics services that you get for the standard metrics captured with the AppDynamics application and machine agents. These services include automatic baselining, anomaly detection, display in the Metric Browser, availability for display on custom dashboards and availability for use in policies to trigger alerts and other actions.
This topic describes the procedure for creating a monitoring extension in Java.
Machine Agent Required
A monitoring extension requires a standalone Machine Agent installed on the machine that hosts the application that you want to monitor.
If you do not know whether you have a Machine Agent installed:
- In the upper right corner of AppDynamics console, click Setup.
- Click AppDynamics Agents.
The list of agents appears with the Machine Agents below the app agents. You can get summary information about the machine and the application associated with each machine agent.
If you do not already have a Machine Agent installed, install one. See Install the Standalone Machine Agent.
To the Machine Agent, a monitoring extension is a task that runs on a fixed schedule and collects metrics. The task can be either AJavaTask, which executes a task within the machine agent process, or AForkedJavaTask, which executes a task in its own separate process.
Before You Begin
Before creating your own extension from scratch, look at the extensions that have been created and shared among members of the AppDynamics community. The extensions are described and their source is available for free download at:
New extensions are constantly being added. It is possible that someone has already created exactly what you need or something close enough that you can download it and use it after making a few simple modifications.
Process for Creating a Monitoring Extension in Java
To create a monitoring extension in Java:
- Create your extension class. See Your Monitoring Extension Class.
- Create a monitor.xml configuration file. See The monitor.xml File.
- Create a zip file containing these two files plus any dependent jar files.
- Create a subdirectory (<your_extension_dir>) in your AppDynamics Machine Agent directory under Machine Agent/monitors.
- Unzip the zip file into Machine Agent/monitors/<your_extension_dir>.
- Restart the Machine Agent.
Your Monitoring Extension Class
Create a monitoring extension class by extending the AManagedMonitor class in the com.singularity.ee.agent.systemagent.api package.
You will also need the following helper classes in the same package:
The Javadoc for these APIs is available at:
Your monitor extension class performs these tasks:
- Populates a hash map with the values of the metrics that you want to add to AppDynamics.
How you obtain these metrics is specific to your environment and to the source from which you derive your custom metrics.
- Describes the metrics using the MetricWriter class.
- Uploads the metrics to the Controller using the execute() method of the TaskOutput class.
All custom metrics processed by the Machine Agent appear in the Application Infrastructure Performance tree in the metric hierarchy.
Within the Application Infrastructure Performance tree you specify the metric path, which is the position of the metric in the metric hierarchy, using the "|" character. The first step in the metric path must be "Custom Metrics."
- Custom Metrics|WebServer|
- Custom Metrics|WebServer|XXX|, CustomMetrics|WebServer|YYY|
If the metrics apply to a specific tier, use the metric path for the tier, with "Component" followed by a colon ":" and the tier ID. The metric will appear under the specified tier in the metric path. For example:
To discover the component ID of a tier, select the tier from the Controller UI application tree. The component ID appears in the URL of the browser.
You can insert a custom metric alongside an existing type of metric. For example, the following declaration causes the custom metric named pool usage to appear alongside the JMX metrics:
- name=Server|Component:18|JMX|Pool|First|pool usage,value="$pool_value
The metric can then be used in health rules as would other types of JMX metrics.
You can test the appearance of your custom metric in the Controller API by posting the metric data to the machine agent's REST API. Pass the path, name type and values of the metric as URL arguments. See Standalone Machine Agent HTTP Listener for more information.
Metric names must be unique within the same metric path but need not be unique for the entire metric hierarchy.
It is a good idea to use short metric names so that they will be visible when they are displayed in the Metric Browser.
Prepend the metric path to the metric name when you upload the metrics to the Controller.
The Controller has various qualifiers for how it processes a metric with regard to aggregation, time rollup and tier rollup. You specify these options with the enumerated types provided by the MetricWriter class. These types are defined below.
The aggregator qualifier specifies how the values reported during a one-minute period are aggregated.
Average of all reported values in the minute. The default operation.
Sum of all reported values in the minute. This operation causes the metric to behave like a counter.
Last reported value in the minute. If no value is reported in that minute, the value from the last time it was reported is used.
Time Roll Up
The time-rollup qualifier specifies how the Controller rolls up the values when it converts from one-minute granularity tables to 10-minute granularity and 60-minute granularity tables over time.
Roll up Strategy
Average of all one-minute data points when adding it to the 10-minute or 60-minute granularity table.
Sum of all one-minute data points when adding it to the 10-minute or 60-minute granularity table.
Last reported one-minute data point in that 10-minute or 60-minute interval.
Cluster Roll Up
The cluster-rollup qualifier specifies how the controller aggregates metric values in a tier.
Roll up Strategy
Aggregates the metric value by averaging the metric values across each node in the tier.
Aggregates the metric value by adding up the metric values for all the nodes in the tier.
For example, if a tier has two nodes, Node A and Node B, and Node A has 3 errors per minute and Node B has 7 errors per minute, the INDIVIDUAL qualifier reports a value of 5 errors per minute and COLLECTIVE qualifier reports 10 errors per minute. INDIVIDUAL is appropriate for metrics such as % CPU Busy where you want the value for each node. COLLECTIVE is appropriate for metrics such as Number of Calls where you want a value for the entire tier.
Sample Monitoring Extension Class
The NGinXMonitor class gets the following metrics from the Nginx Web Server and adds them to the metrics reported by AppDynamics:
- Active Connections: number of active connections
- Accepts: number of accepted requests
- Handled: number of handled requests
- Requests: total number of requests
- Reading: number of reads
- Writing: number of writes
- Waiting: number of keep-alive connections
Here is the source for the extension class.
The monitor.xml File
Create a monitor.xml file with a <monitor> element to configure how the machine agent will execute the extension.
- Set the <name> to the name of your Java monitoring extension class.
- Set the <type> to "managed".
- The <execution-style> can be "continuous" or "periodic".
- Continuous means to collect the metrics averaged over time; for example, average CPU usage per minute. In continuous execution, the Machine Agent invokes the extension once and the program runs continuously, returning data every 60 seconds.
- Periodic means to invoke the monitor at a specified frequency. In periodic execution the Machine Agent invokes the extension, runs it briefly, and returns the data on the schedule set by the <execution-frequency-in-seconds> element.
- If you chose "periodic" for the execution style, set the frequency of collection in <execution-timeout-in-secs> element. The default frequency is 60 seconds. If you chose "continuous" this setting is ignored.
- Set the <type> in the <monitor-run-task> child element to "java".
- Set the <execution-timeout-in-secs> to the number of seconds before the extension times out.
- Specify any required task arguments in the <task-arguments> element. The default arguments that are specified here are the only arguments that the extension uses. They are not set anywhere else.
- Set the <classpath> to the jar file that contains your extension's classes. Include any dependent jar files, separated by semicolons.
- Set the <impl-class> to the full path of the class that the Machine Agent invokes.
Sample monitor.xml Files
The following monitor.xml file configures the NGinXMonitor monitoring extension. This extension executes every 60 seconds.
The next monitor.xml file configures the MysqlMonitor. This monitor executes every 60 seconds, has four required task arguments and one optional task argument and one dependent jar file.
Using Your Monitoring Extension as a Task in a Workflow
Your monitoring extension can be invoked as a task in a workflow if you upload the zip file to the task library. Use the instructions in To package the XML files as a Zip archive to upload the Java monitor to the Task Library.