Foglight for Container Management 2.0 - User Guide

Rules

Foglight for Container Management allows you to create flexible rules that can be applied to complex interrelated data from multiple sources within your clusters. You can associate several different actions with a rule, configure a rule so that it does not fire repeatedly, and associate a rule with schedules to define when it should be evaluated or not.

Different types of data can be used in rules, including registry variables, raw metrics, derived metrics, and topology object properties.

There are two types of rules: simple rules and multiple-severity rules. A simple rule has a single condition, and can be in one of three states: Fire, Undefined, or Normal. A multiple-severity rule can have up to five severity levels: Undefined, Fatal, Critical, Warning, and Normal.

Rule conditions are regularly evaluated against monitoring data (metrics and topology object properties collected from your monitored environment and transformed into a standard format). Therefore, the state of the rule can change if the data changes. For example, if a set of monitoring data matches a simple rule’s condition, the rule enters the Fire state. If the next set does not match the condition, the rule exits the Fire state and enters the Normal state.

Rules can be configured to send emails, pager messages, or perform other actions you define. Performance data can be viewed and analyzed using Foglight for Container Management.

Foglight for Container Management includes a number of predefined rules used to monitor the health of your container clusters. You are allowed to modify these rules to satisfy your different requirements. Many of these rules listed and described in this section have thresholds defined within them. Those thresholds include standard deviations, utilization percentages, and so on, are default values predefined in the registry.

Kubernetes

All rules are controlled by registry variable Kubernetes:AlertSensitivity. If the value is 0, then no alarm can be fired. If the value is 1, warning level alarm can be fired. If the value is above 1, then all level alarm can be fired.

Kubernetes Administrator email address can be configured in Registry Variable KubernetesAdmin.

Health Check

This rule detects abnormal Pod health status and fires alarm for different severity abnormal health status.

KubePod

This rule detects Pods that stays in pending phase for an abnormal long time.

KubePod

This rule detects abnormal Container health status and fires alarm for different severity abnormal health status.

KubeContainer

This rule detects abnormal Node health status and fires alarm for different severity abnormal health status.

KubeNode

This rule detects abnormal Deployment health status and fires alarm for different severity abnormal health status.

KubeDeployment

This rule detects abnormal Daemon Set health status and fires alarm for different severity abnormal health status.

KubeDaemonSet

This rule detects abnormal Job health status and fires alarm for different severity abnormal health status.

KubeJob

This rule detects abnormal Persistent Volume health status and fires alarm for different severity abnormal health status.

KubePersistentVolume

This rule detects abnormal Persistent Volume Claim health status and fires alarm for different severity abnormal health status.

KubePersistentVolumeClaim

This rule detects abnormal long pending Persistent Volume Claim and fires alarm for different severities.

KubePersistentVolumeClaim

Usage

This rule detects abnormal CPU Utilization for Pods, and fires alarm on different severities. It is disabled by default. You can customize it and enable it based on your different requirements. For more details about customization, refer to Customization . You can also change value of the registry variables or use your own value to change the threshold of each severities. This rule only works for those Pods that configures CPU limit.

KubePod.metrics

This rule detects abnormal Memory Utilization for Pods, and fires alarm on different severities. It is disabled by default. You can customize it and enable it based on your different requirements. For more details about customization, refer to Customization . You can also change value of the registry variables or use your own value to change the threshold of each severities. This rule only works for those Pods that configures Memory limit.

KubePod.metrics

This rule detects abnormal CPU Usage for Pods, and fires alarm on different severities. It is disabled by default. You can customize it and enable it based on your different requirements. For more details about customization, refer to Customization . You can also change value of the registry variables or use your own value to change the threshold of each severities.

KubePod.metrics

This rule detects abnormal Memory Usage for Pods, and fires alarm on different severities. It is disabled by default. You can customize it and enable it based on your different requirements. For more details about customization, refer to Customization . You can also change value of the registry variables or use your own value to change the threshold of each severities.

KubePod.metrics

This rule detects abnormal Network Receive in bytes for Pods, and fire alarm on different severities. It is disabled by default. You can customize it and enable it based on your different requirements. For more details about customization, refer to Customization . You can also change value of the registry variables or use your own value to change the threshold of each severities.

KubePod.metrics

This rule detects abnormal Network Send in bytes for Pods, and fire alarm on different severities. It is disabled by default. You can customize it and enable it based on your different requirements. For more details about customization, refer to Customization . You can also change value of the registry variables or use your own value to change the threshold of each severities.

KubePod.metrics

This rule detects abnormal CPU Utilization in bytes for Nodes, and fire alarm on different severities. It is enabled by default. You can change value of the registry variables or use your own value to change the threshold of each severities.

KubeNode.metrics

This rule detects abnormal Memory Utilization in bytes for Nodes, and fire alarm on different severities. It is enabled by default. You can change value of the registry variables or use your own value to change the threshold of each severities.

KubeNode.metrics

This rule detects abnormal Network Receive in bytes for Nodes, and fire alarm on different severities. It is enabled by default. You can change value of the registry variables or use your own value to change the threshold of each severities.

KubeNode.metrics

This rule detects abnormal Network Send in bytes for Nodes, and fire alarm on different severities. It is enabled by default. You can change value of the registry variables or use your own value to change the threshold of each severities.

KubeNode.metrics