Skip to main content

Set up

note

Currently, this feature is behind the feature flag SRM_ET_EXPERIMENTAL. Contact Harness Support to enable the feature.

This topic describes how to add a monitored service in Harness and install an Error Tracking Agent on a JVM to identify and troubleshoot faults and exceptions in your code.

Add a monitored service

A Harness Monitored Service is a combination of service and environment. Harness monitors the following via a monitored service:

  • Code-level exceptions and errors from the data collected by the Error Tracking Agent.
  • Changes such as deployments, incidents such as PagerDuty, and infrastructure changes such as Kubernetes events and auditing.
  • Health trend deviations using logs and metrics obtained from the APM and logging tools.

The following steps explain how to create a monitored service, and define service and environments for tracking code errors and exceptions:

  1. In your Harness project, navigate to the Service Reliability Management module, and then select Monitored Services.

    Navigate to Monitored Services

  2. Select + New Monitored Service.

    The Create new monitored service settings page appears.

  1. In the Overview section, enter the following information:

    • Type: Select Application.

    • Create or select a Service: Select a Harness service that you want to monitor. A Harness service is a logical representation of your microservices and other workloads that you want to monitor.

      To create a new service, perform the following steps:

      1. Select + Add New.

        The New Service dialog appears.

      2. Enter the following information, and then select Save:

        • Name : Enter a name for the service. For example, sample_service.
        • Description (Optional)
        • Tag (Optional)
    • Create or Select an Environment: Choose an appropriate environment. Environments represent your deployment targets such as QA, Prod, and so on.

      To create a new environment, perform the following steps:

      1. Select + Add New.

        The New Environment dialog appears.

      2. Enter the following information and then select Save:

        • Name: Enter a name for the environment.
        • Description (Optional)
        • Tag (Optional)
        • Environment Type: Select an environment type. The available options are Production and Non-Production.

    Harness autopopulates the Monitored Service Name field by combining the Service and Environment names.

  2. Select Save.

    You can view the new monitored service in the Monitored Services list.

    Monitored Service is created

Create a token for the Error Tracking Agent

To create a token, perform the following steps:

  1. Expand PROJECT SETUP and then select Code Error Settings.

  2. On the top right-hand corner, select Tokens.

    Create Token

  3. Select Generate New Token.

    The New Token dialog appears.

  4. Enter a Name for the token.

  5. Select Generate. A new authentication token is generated and displayed in the Key field.

  6. Copy the token to clipboard. This token is used in the Error Tracking Agent configuration file.

    Generate new token

  7. Select Close.

    The newly created token is added to the list of tokens.

New token

note

You can copy the token anytime by selecting the clipboard icon.

Install an Error Tracking Agent

The Harness Error Tracking Agent must be installed on a Java application to capture application code and variables. This native agent library is attached to the Java Virtual Machine (JVM)/Common Language Runtime (CLR) during runtime. It can be deployed into any Kubernetes container without altering the image. The Agent fingerprints the program code loaded into the JVM/CLR and captures the complete state of the code and the host/container in order to identify anomalies.

This topic provides details about:

  • Compatibility and requirements to install an Error Tracking Agent.
  • Steps to install an Error Tracking Agent on a Java Application.

Compatibility and Requirements

This table lists the supported operating system, Java Virtual Machine (JVM) versions, and JVM containers.

Future releases will include support for more operating systems.

Operating SystemSupported JVM VersionsSupported JVM Containers
Linux Operating System:
  • Ubuntu: 14+
  • jDebian
  • CentOS: 6.5+
  • RedHat: 5.0 +
  • Suse: SLES12
  • Oracle/HotSpot:6u20 - 6u457 - 7u808 - 8u2218 - 8u2329 - all updates10/11/16/17 - all updates
  • OpenJDK:6u20 - 6u457 - 7u808 - 8u2228 - 8u2329 - all updates10/11/16/17 - all updates
  • Jetty
  • Scala
  • Eclipse
  • NetBeans
  • IntelliJ
  • JBoss/Wildfly
  • CloudFoundry
  • Weblogic
  • Play Framework
  • Glassfish
  • Mule
  • WebSphere
  • Tomcat
JVM Requirements

When you attach the Harness Error Tracking Agent to a JVM that runs Java 10, 11, 16, 17, or any IBM Java version, ensure that the following requirements are met:

  • Turn off class sharing using the following flags:

    IBM Java‑Xshareclasses:none
    HotSpot-Xshare:off -XX:-UseTypeSpeculation
  • Increase ReservedCodeCache to at least 512mb by adding the following flag:
    -XX:ReservedCodeCacheSize=512m

Install an Error Tracking Agent

Depending on your setup and needs, there are multiple ways to install and deploy an Error Tracking Agent on a Java application. You can install the Agent to monitor your application in the following ways:

  • As a standalone installation outside Harness.
  • By modifying your Docker image.
  • Using an init container.

If you have additional agents running on your application, the Error Tracking Agent should appear at the end of the VM arguments list before specifying the main class or jar.

This option lets you install the Error Tracking Agent as a standalone. Perform the following steps to install the Agent:

  1. Download the latest version of the Agent for Linux.

  2. Extract this version to a folder of your choice. For example/home/user. The contents of the archive will be inside a folder named Harness. You'll require the path to the Agent when starting your application. An example for the Agent path would be , /home/user/harness/lib/libETAgent.so.

  3. Add JVM arguments which instruct the JVM to load the Agent.

    This is done by adding -agentpath:/home/user/harness/lib/libETAgent.so to the application startup. For example:

    java -agentpath:/home/user/harness/lib/libETAgent.so -jar yourapp.jar.

    This parameter can also be specified using JAVA_TOOL_OPTIONS. For example:

    exportJAVA_TOOL_OPTIONS=-agentpath:/home/user/harness/lib/libETAgent.so.

  4. Set the Agent environment variables so that your application can map to a Harness Service.

    Required Environment VariableDescriptionExample
    ET_COLLECTOR_URLURL to the Error Tracking collector.https://collector.et.harness.io/prod1
    ET_APPLICATION_NAMEName of your application or Service.myapp
    ET_DEPLOYMENT_NAMEDeployment or version number of your application or Service. When your application or Service is updated to a new version, it's recommended that you update this variable as well, so that the Error Tracking Agent can identify when new errors are introduced.1
    ET_ENV_IDID of your Harness Environment.production
    ET_TOKENET Agent Token created on Harness.b34a3f1a-7b38-4bb6-b5fe-49f52314f5342a

    For example:

ENV ET_COLLECTOR_URL=https://collector.et.harness.io/prod1/
ENV ET_APPLICATION_NAME=yourapp
ENV ET_DEPLOYMENT_NAME=1
ENV ET_ENV_ID=env1
ENV ET_TOKEN= agenttoken
  1. Restart your application after installing the Error Tracking Agent.

Verify Agent connection

After installing Error Tracking Agent, you should verify that it is connected to Harness SRM. To verify the Error Tracking Agent connection, do the following:

  1. Expand PROJECT SETUP and then select Code Error Settings.

    Navigate to Agents list

    A comprehensive list of Error Tracking Agents is displayed along with the information such as service name, environment, deployment version, Agent version, Agent status, token name, and so on. Ensure that the Agent that you installed is listed, and the status is CONNECTED.

    Verify agent connection

Next steps

Identify and prioritize error events using Events dashboard.