K8s Cluster

Overview

K8s Cluster collects cluster-level state metrics from the Kubernetes API server. It provides visibility into node conditions, resource allocation, workload states, and metadata across your cluster. It is ideal for monitoring cluster health, scheduling behavior, and infrastructure-level signals.

Supported platforms

• Kubernetes: Metrics

Workload Configuration

Parameter	Type	Default	Required	Description
kind	string	Deployment	Yes	The source runs as a Deployment.
enable_leader_election	boolean	true	No	Ensures only one pod scrapes the API at a time (HA mode).
replica_scaling_type	string	manual	No	Scaling mode: manual or auto.
replicas	integer	2	Conditional	Required when replica_scaling_type = manual.
priorityClass	boolean	false	No	Assign priority class to pods.
priorityClassName	string	—	Conditional	Required when priorityClass = true.
apply_tolerations	boolean	false	No	Allow scheduling on control-plane nodes.

Metric & Metadata Configuration

Parameter	Type	Default	Required	Description
collection_interval	string	10s	No	Frequency at which cluster metrics are collected.
resource_attributes	array	["container.id", "k8s.pod.uid"]	No	Additional resource attributes attached to metrics.
node_conditions_to_report	array	["Ready"]	No	Node conditions emitted as metrics (supported options: Ready, MemoryPressure, DiskPressure, PIDPressure, NetworkUnavailable).
allocatable_types_to_report	array	[]	No	Node allocatable resources to track (options: cpu, memory, pods, ephemeral-storage).

Advanced Settings

Parameter	Type	Default	Required	Description
distribution	string	kubernetes	No	Kubernetes distribution (kubernetes or openshift).
namespaces	array	[]	No	Limit metrics to specific namespaces.
metadata_collection_interval	string	5m	No	Interval for refreshing metadata.

Example Configuration

{
  // Workload Configuration
  "kind": "Deployment",
  "enable_leader_election": true,
  "replica_scaling_type": "manual",
  "replicas": 2,
  "priorityClass": false,
  "priorityClassName": "",
  "apply_tolerations": false,

  // Metric & Metadata Configuration
  "collection_interval": "10s",
  "resource_attributes": ["container.id", "k8s.pod.uid", "k8s.node.uid"],
  "node_conditions_to_report": ["Ready", "MemoryPressure", "DiskPressure"],
  "allocatable_types_to_report": ["cpu", "memory", "pods"],

  // Advanced Settings
  "distribution": "kubernetes",
  "namespaces": ["production", "staging"],
  "metadata_collection_interval": "5m",
}