Get the latest, first
Leveraging Golden Signals for Enhanced Kubernetes Security

Leveraging Golden Signals for Enhanced Kubernetes Security

Jun 20, 2024

Oshrat Nir
Developer Advocate

As a powerful and widely adopted open-source platform, the complexity of Kubernetes is not to be underestimated. Managing a Kubernetes environment requires a deep understanding of how its various components interact, especially when it comes to observability and security. 

This blog post will delve into the intricacies of golden signals in Kubernetes, their connection to security issues, and how they can be leveraged to safeguard a Kubernetes environment against common attack chains.

Understanding Golden Signals

Originating from the site reliability engineering (SRE) discipline, golden signals are a set of key performance indicators (KPIs) that are essential for monitoring and managing the user experience and system reliability. They act as an early warning system that identifies issues with the system’s performance, security, and availability objectives.

Importance for Modern DevOps/SRE Teams

Technology today is fast-paced and ever-changing, making the role of DevOps/SRE teams more crucial than ever. These teams ensure seamless application integration, deployment, and operation, often across diverse and dynamic environments. Golden signals are indispensable tools in their toolbox, enabling teams to:

  • Proactively identify issues: Golden signals help detect performance bottlenecks, system overloads, and failures early on, allowing teams to address them before they impact users.
  • Optimize system performance: By analyzing golden signals, teams can identify optimization opportunities, allocate resources efficiently, and enhance overall system performance.
  • Enhance user experience: Monitoring and optimizing based on golden signals contribute to improved application responsiveness, availability, and user satisfaction.
  • Strengthen security posture: Anomalies in golden signals can indicate potential security threats, enabling timely detection and mitigation of vulnerabilities and attacks.

The Four Primary Golden Signals

Four main golden signals form the cornerstone of system monitoring and observability.

1. Latency

Latency tells you how long it takes for a request to be processed. It is a critical indicator of system responsiveness. High or increasing latency can signal performance bottlenecks, resource contention, or potential security threats. 

For instance, in the realm of security, a sudden spike in latency could be indicative of a distributed denial-of-service (DDoS) attack, where attackers flood a system with traffic to make it slow or unavailable. Monitoring latency helps identify and mitigate such threats promptly.

2. Traffic

Traffic represents the number of requests received by the system. Monitoring this metric helps understand system load, identify usage patterns, and detect anomalies such as sudden traffic spikes, which could indicate a DDoS attack or a popular feature. 

For example, a significant increase in traffic from a specific geographical location, especially one not typically associated with your user base, could indicate your system is being hit by a coordinated cyberattack or an attempt to probe vulnerabilities. As a result, recognizing this unusual traffic pattern early enables teams to initiate appropriate defensive measures.

3. Errors

Errors indicate the rate of failed requests. A high error rate can mean there are problems with the application code, system resources, or dependencies, or you may have potential security vulnerabilities. Consider this scenario: An online banking application starts registering an unusual error rate specifically for fund transfers between accounts. Upon investigation, you discover that these errors are due to a new software update that inadvertently introduced a bug in the transaction validation process. If not caught and remediated immediately, this could lead to unauthorized transfers, potential financial loss, and a severe dent in user trust.

4. Saturation

Saturation measures the utilization of resources such as CPU and memory. High saturation levels indicate the system is nearing capacity, affecting performance and potentially leading to failures. 

Let’s take a look at this from a security perspective. Say a container within a Kubernetes cluster experiences sudden, high saturation levels in its CPU utilization. Further examination finds that malicious software is running a crypto-mining operation within the container, using up its resources. Monitoring saturation allows the security team to detect the breach and initiate the necessary countermeasures.

In the context of Kubernetes security, all four of these signals are pivotal in detecting and mitigating security risks in containerized environments.

Golden Signals in Kubernetes: A Security View

As a complex and dynamic orchestration platform, Kubernetes presents unique challenges and opportunities for implementing and monitoring golden signals. This section will explore how the four primary golden signals can be contextualized in Kubernetes and connected to potential security issues.

1. Latency 

In Kubernetes, latency is the time taken for a request to travel from a service to a pod. High latency can indicate network issues, resource contention, or anomalies. Unusual spikes in latency can indicate potential threats such as network probing, data exfiltration, or resource exhaustion attacks. 

2. Traffic 

Traffic in Kubernetes is represented by the number of requests sent to a service or a pod. Monitoring traffic helps identify usage patterns and detect abnormal spikes or drops. A sudden and unexplained increase in traffic can be a sign of a DDoS attack, unauthorized access attempts, or potential security breaches.

3. Errors 

Errors in a Kubernetes environment can manifest as a pod failing to respond to a request, a service failing to route a request, or a pod crashing. A high error rate can signal vulnerabilities, misconfigurations, or active exploitation attempts.

4. Saturation

Saturation in Kubernetes is measured by monitoring a node or pod’s CPU or memory usage. It shows how close resources are to their capacity. For instance, excessive resource consumption or high saturation levels can be indicators of compromised workloads, cryptomining attacks, or resource exhaustion attacks.

By contextualizing latency, traffic, errors, and saturation within the Kubernetes environment and understanding their security implications, organizations can enhance their security posture and protect their containerized applications and infrastructure from a myriad of cyber threats. 

Role of Golden Signals in Detecting Attack Chains

In the realm of Kubernetes security, understanding the common attack chains is pivotal. Golden signals play a significant role in their early detection, allowing organizations to thwart potential breaches. Below, we explain how monitoring golden signals can help detect the top four attack chains discussed in our earlier post.

1. Exposed Endpoint Attacks

  • Traffic: A sudden spike in traffic to exposed endpoints can indicate a potential attack.
  • Errors: An increase in error rates may signify unauthorized access attempts.

2. Privilege Escalation Attacks

  • Latency: High latency in service-to-pod requests may indicate unauthorized access attempts.
  • Errors: Elevated error rates can signify issues with RBAC policies or unauthorized access.

3. Supply Chain Attacks

  • Traffic: Unusual traffic patterns to or from container registries can indicate compromised images.
  • Saturation: Increased resource usage may signal the deployment of malicious workloads.

4. Developer Credential Theft

  • Latency: Increased latency in accessing resources can indicate unauthorized access using stolen credentials.
  • Traffic: Unusual traffic patterns, such as increased requests to Git repositories, can signal potential credential theft.

Best Practices for Monitoring Kubernetes Golden Signals

As with everything, when using golden signals in your Kubernetes environment, there are some actions any organization should take. 

Set Up Alerts

One of the foundational steps in monitoring Kubernetes golden signals is to create alerts. Alerts act as your frontline defense, notifying teams of potential issues before they escalate. However, it’s important to be mindful of alert fatigue. Having too many alerts – important ones, mixed with non-critical ones can overwhelm teams, and divert their attention away from truly significant issues. It’s critical to ensure that the alerts you set up are meaningful and actionable.

Suggested alerts:

  • Threshold-based alerts: Define clear thresholds for each golden signal. For instance, set alerts for when latency exceeds a specific limit, error rates spike, or resource saturation reaches a critical level.
  • Anomaly detection: Implement anomaly detection to identify unusual patterns or deviations in traffic, latency, errors, and saturation, which could indicate potential security incidents.
  • Contextual alerts: Customize alerts based on the specific characteristics and requirements of different workloads and services within the Kubernetes cluster.
  • Actionable alerts: Ensure that alerts provide sufficient information and context to enable quick diagnosis and remediation of any identified issues.

Review Metrics Regularly 

Continuously revisiting your golden signal metrics is essential for maintaining a proactive stance on Kubernetes security. It’s not just about reacting to alerts but also analyzing trends, identifying areas for improvement, and anticipating potential issues. Here are some strategies for effective metric reviews:

  • Trend analysis: Analyze the historical data of golden signals to identify trends, patterns, and possible anomalies.
  • Performance baselines: Establish and regularly update performance baselines for each service and workload to quickly identify deviations and anomalies.
  • Capacity planning: Use saturation metrics to inform capacity planning and resource allocation decisions, ensuring optimal performance and avoiding resource exhaustion.
  • Security correlation: Correlate golden signal metrics with security logs and other indicators to identify and investigate potential security incidents.

Use Tools for Monitoring

Leveraging the right tools is crucial for effectively monitoring golden signals in Kubernetes. Several popular ones can aid in this endeavor:

  • Prometheus: A widely used and open-source toolkit for reliability and scalability, Prometheus is particularly well-suited for monitoring Kubernetes clusters, providing insights into latency, traffic, errors, and saturation.
  • Grafana: Also open-source, Grafana integrates seamlessly with Prometheus, featuring dashboards and visualization tools for analyzing Kubernetes golden signals.
  • Jaeger: A distributed tracing system that can trace requests across services in Kubernetes, Jaeger offers insights into latency and helps identify performance bottlenecks.

By utilizing these tools, teams can comprehensively view the Kubernetes environment, monitor golden signals effectively, and respond promptly to any anomalies or security threats.

Golden Signals and Security

The symbiosis between golden signals and security is evident. Anomalies in latency, traffic spikes, elevated error rates, and resource saturation can all be indicators of security incidents. Whether it’s a misconfiguration exploit, a container breakout attempt, or a privilege escalation attack, golden signals act as an early warning system, enabling timely detection and mitigation.

By setting up contextual alerts, conducting regular metric reviews, and utilizing monitoring tools, organizations can harness the power of golden signals to safeguard their Kubernetes clusters. As a result, this proactive approach to monitoring and security is indispensable in today’s cyber threat landscape.


Golden signals are the linchpins of system observability and performance monitoring. They offer invaluable insights into the health and functioning of applications and infrastructure, enabling teams to proactively uncover and remediate issues. In addition, they serve as sentinels, guarding against performance bottlenecks, system failures, and security vulnerabilities.

How can ARMO Platform help?

ARMO Platform is a powerful solution designed to examine and enhance the security posture of Kubernetes clusters. It performs comprehensive assessments, identifies misconfigurations, and provides actionable recommendations to fortify cluster security.

ARMO Platform can use Kubernetes golden signals to bolster security in several ways:

  • Comprehensive assessments: It thoroughly assesses Kubernetes clusters, identifying vulnerabilities, misconfigurations, and deviations from best practices.
  • Actionable recommendations: The platform provides detailed recommendations and remediation steps to enhance cluster security when an issue is identified.
  • Continuous monitoring: It supports the monitoring of Kubernetes environments, enabling real-time detection of security risks and compliance drifts.
  • Integration with monitoring tools: The platform can be integrated with popular monitoring solutions for enhancing the visibility and analysis of golden signals.

By leveraging ARMO Platform, organizations can gain deeper insights into their enhanced Kubernetes security posture, proactively address vulnerabilities, and ensure the resilience and integrity of their containerized applications and infrastructure.

Start using ARMO Platform today, and learn how to utilize golden signals to boost the security of your Kubernetes environment.

Actionable, contextual, end-to-end
{Kubernetes-native security}

From code to cluster, helm to node, we’ve got your Kubernetes covered:

Cut the CVE noise by significantly reducing CVE-related work by over 90%

Automatic Kubernetes compliance for CIS, NSA, Mitre, SOC2, PCI, and more

Manage Kubernetes role-based-access control (RBAC) visually


Continue to Slack

Get the information you need directly from our experts!

new-messageContinue as a guest