Spring Boot 3: Disable Metrics

Effortlessly disable metrics in Spring Boot 3. Learn simple steps to control and exclude metrics for a streamlined application.

Introduction

Spring boot 3 provides a powerful and flexible way to collect and expose various metrics for your application, such as JVM, CPU, memory, HTTP, cache, etc. However, sometimes you may want to disable some or all of the metrics including the custom metrics that are defined in your application. In this blog post, we will learn how to do that using the application properties.

Decoding Spring Boot Actuator Metrics

In Spring Boot, metrics work as performance indicators, giving crucial insights into your app’s operation. They cover CPU usage, memory, HTTP requests, and more. Spring Boot 3 makes monitoring easy, letting devs track default metrics and make custom ones effortlessly. These metrics are essential, helping find improvements, optimize resources, and keep your app performing at its peak.

Spring boot 3 uses Micrometer, an application metrics facade, to support various monitoring systems such as Prometheus, Datadog, Graphite, etc. Micrometer allows you to create and register custom metrics using its fluent API. For example, you can create a counter to measure the number of requests to a certain endpoint, or a timer to measure the latency of a method call. You can also add tags to your metrics to provide more dimensions and context.

When and Why Exclude Metrics from Actuator Metrics Endpoint

In certain scenarios, excluding specific metrics from the actuator metrics endpoint becomes advantageous. For instance:

• Metrics that are not relevant or useful for your application or monitoring system. For example, you may not need the metrics for the cache system if you are not using any caching in your application.
• Metrics that are redundant or overlapping with other metrics. For example, you may not need the metrics for the HTTP server requests if you are already using a custom metric to measure the same aspect of your application.
• Metrics that are sensitive or confidential and may pose a security or privacy risk. For example, you may not need the metrics for the JPA system if they contain information about your database queries or entities that may expose your data or business logic.

How to Disable All Metrics?

In some rare cases, you may want to disable all metrics, including your custom metrics, from being exposed or exported. For example, you may want to reduce the resource consumption or the network traffic of your application or monitoring system, or you may want to avoid any potential conflicts or errors caused by the metrics. In this case, you have two options:

1. Use the management.metrics.enable.all property to false

management.metrics.enable.all=false

This will disable all metrics, including your custom metrics, from being exposed or exported.

After adding this configuration, the metrics endpoint will be accessible but it will return an empty array meaning no metrics information will be available in this case.

2. Exclude MetricsAutoConfiguration.class by specifying it inside the @SpringBootApplication annotation.

@SpringBootApplication(exclude = MetricsAutoConfiguration.class)
public class Application {

  public static void main(String[] args) {
    SpringApplication.run(Application.class, args);
  }
}

This will disable the MetricsAutoConfiguration class, which is the main auto-configuration class for metrics in Spring Boot 3.

It is important to note that after excluding the MetricsAutoConfiguration class, attempting to access the Actuator metrics endpoint will result in a 404 error, as the metrics will no longer be accessible.

Disable Default and Custom Metrics

How to Disable Default Metrics?

Spring Boot 3 also provides a number of default metrics for common aspects of your application, such as JVM, CPU, memory, HTTP, cache, etc. These metrics are configured via various auto-configuration classes that are enabled by default. 

To disable the default metrics, you have two options:

• Disable all metrics and enable only specific metrics
• Disable only selected metrics and keep the rest

Option 1: Disable all metrics and enable only specific metrics

To disable all metrics and enable only specific metrics, you need to set the management.metrics.enable.all property to false and the management.metrics.enable.<metric> property to true in your application.properties file. For example:

management.metrics.enable.all=false
management.metrics.enable.jvm=true

This will disable all the default metrics and only expose metrics that start with jvm.

Option 2: Disable only selected metrics and keep the rest

To disable only selected metrics and keep the rest, you need to set the management.metrics.enable.<metric> property to false for each metric that you want to disable in your application.properties file. The <metric> is the name of the metric that you want to disable. For example, to disable the JVM, CPU, and memory metrics, you can use the following properties:

management.metrics.enable.jvm=false
management.metrics.enable.cpu=false
management.metrics.enable.memory=false

This will disable the JVM, CPU, and memory metrics and keep the rest of the default metrics.

How to Disable Custom Metrics?

Sometimes, you may want to exclude some of your custom metrics from being exposed or exported. For example, you may have some metrics that are only relevant for development or testing purposes or some metrics that are too sensitive or confidential to be exposed. In this case, you can use the management.metrics.enable.<metric> property to false for each custom metric that you want to exclude. For example, to exclude the myapp.secret metric, you can use the following property:

management.metrics.enable.myapp.secret=false

This will exclude the myapp.secret metric from being exposed or exported.

Disable Metrics Using Patterns

Sometimes, you may want to disable a group of metrics that share a common prefix in their names. For example, you may want to disable all the metrics that start with jvm, such as jvm.buffer.count, jvm.gc.live.data.size, jvm.threads.daemon, etc. In this case, you can use the management.metrics.enable.<pattern> property to false for the pattern that you want to disable. The <pattern> is the prefix of the metric names that you want to disable. For example, to disable all the metrics that start with jvm, you can use the following property:

management.metrics.enable.jvm=false

This will disable all the metrics that start with jvm. You can use this property for any pattern that matches the metric names. For example, to disable all the metrics that start with executor.pool, such as executor.pool.core, executor.pool.max, executor.pool.size, etc., you can use the following property:

management.metrics.enable.executor.pool=false

This will disable all the metrics that start with executor.pool.

Things to Consider

Before you disable or exclude any metrics, you should consider the following points:

• Functional Impact: Disabling or excluding metrics may affect the functionality or performance of your application or monitoring system. For example, some metrics may be used by health indicators, alerts, dashboards, or other features that rely on them. Make sure you understand the implications of disabling or excluding any metrics before you do so.
• Retained Metrics Alert: Disabling or excluding metrics may not completely remove them from your application or monitoring system. For example, some metrics may still be registered in the MeterRegistry or stored in the monitoring system’s database, even if they are not exposed or exported. This may consume some resources or cause some errors or warnings. To completely remove any metrics, you may need to use other methods, such as customizing the MeterRegistry or deleting the metrics from the monitoring system.
• Data Privacy Caution: Disabling or excluding metrics may not be enough to protect your data or privacy. For example, some metrics may contain sensitive or confidential information, such as user names, passwords, tokens, etc. Even if you disable or exclude these metrics, they may still be accessible or visible to some parties, such as developers, administrators, or hackers. To protect your data or privacy, you may need to use other methods, such as encryption, masking, or anonymization.

FAQs

Conclusion

In summary, selectively excluding certain metrics from Spring Boot’s actuator endpoint is a smart move. It helps streamline monitoring by removing unnecessary, duplicate, or sensitive metrics. By doing this, you ensure that the displayed metrics perfectly match your monitoring requirements, making your system more efficient and secure.

Learn More