Introduction to New Relic for Java

With New Relic's Java agent, you can track everything from performance issues to tiny errors within your code. Every minute the agent posts metric timeslice and event data to the New Relic user interface, where the owner of that data can sign in and use the data to see how their website is performing.

JSR 163 compliance

The New Relic Java agent monitors web transactions, receiving information about them and communicating that information to the New Relic user interface with strong data security measures in place. The agent is packaged as a JSR 163 compliant javaagent that is activated by the JVM through modifications to the JVM launch. Once activated, the agent inserts itself into the class loading stream and instruments class methods using byte code instrumentation (bci).

Designed to have minimal impact on your web application, all of the classes are in the newrelic package namespace so they do not collide with your own classes. The agent uses the ASM bci engine to insert software probes.

The agent receives basic information about your host environment, such as operating system, Java version, system properties, and your New Relic configuration file. The agent also polls data from the JVM and from JMX.

Installation

To use the Java agent:

  1. Make sure your system meets the Java agent's compatibility requirements.
  2. Sign up for a New Relic account.
  3. Install the Java agent by following the standard installation procedures. Depending on your tools and frameworks, refer to additional installation procedures to install or configure the Java agent.
  4. To view your app's performance in the New Relic UI: Go to rpm.newrelic.com/apm > (select an app) > Monitoring > Overview.

The APM user interface includes a dedicated JVM metrics page, transaction and error details, a thread profiler tool to sample Java threads and report stack traces, and more. You can also extend your agent instrumentation and explore your data with other New Relic products, including New Relic Insights.

Startup configuration

The Java agent reads the startup configuration from the newrelic.yml file, which must be in the same directory as newrelic.jar. You can also override settings using Java system properties.

The agent needs the license_key and app_name settings at startup (bootstrap). You can customize the location of the log file and other settings in your newrelic.yml file. In addition, you can adjust the naming configuration if your app server runs multiple applications in the same JVM.

Configuration options

You can configure application reporting from the New Relic web interface using server-side configuration. Changes will be applied to all agents reporting for your application when the JVM restarts. However, if you use server-side configuration, you must still maintain required settings (license_key and app_name) in the local config file.

Both transaction traces and error snapshots can optionally record HTTP parameters. Sometimes HTTP parameters contain sensitive information, such as a credit card number. The agent has several configuration options to ensure data security, such as collecting HTTP parameters but excluding certain named parameters.

Additional instrumentation

After installing the Java agent, use any of these methods to extend your instrumentation:

When to restart your JVM

When you make a change to your configuration or to your custom XML, in almost all cases you must restart your JVM. There are only four situations where a restart is not necessary:

JVM restart is not required for... Comments
Log level or audit mode changes

If you are simply changing the log_level or audit_mode in the Java agent configuration file, you do not need to restart the JVM.

Custom instrumentation XML file

JVM restart depends on your Java agent version:

  • 3.15.0 or higher: If you are updating a custom instrumentation XML file in the extensions folder, the updates will be reloaded automatically within a minute.
  • Java agent versions lower than 3.15.0: If you using an agent version earlier than 3.15.0, you must restart the JVM.
Circuit breaker changes

If you are changing the circuit breaker for Java custom instrumentation, a JVM restart is not required.

Ignored or expected errors

If you are ignoring or expecting errors by configuring the newrelic.yml file, a JVM restart is not required. This includes adding status codes or classes to the newrelic.yml file when using the Java agent version 3.40.0 or higher. For example:

  • ignore_status_codes
  • expected_status_codes
  • ignore_classes
  • ignore_messages
  • expected_classes
  • expected_messages

All other changes require an agent restart, which means restarting the JVM.

Troubleshooting procedures

If you encounter problems with the Java agent, see the troubleshooting documentation.

For more help

Recommendations for learning more: