Skip to content

Latest commit

 

History

History
148 lines (132 loc) · 4.91 KB

configuration.md

File metadata and controls

148 lines (132 loc) · 4.91 KB
Raw

Configuring and Tuning Cluster Logging

OKD cluster logging is configurable by modifying the CustomResource deployed in the openshift-logging namespace. The sample resource defines a complete cluster logging deployment that includes all the subcomponents of the logging stack to collect, store and visualize logs. The cluster-logging-operator watches the a ClusterLogging CustomResources and adjusts the logging deployment accordingly.

Note: The cluster-logging-operator will only handle a ClusterLogging instance named instance. This is to enforce a single ClusterLogging deployment for the entire OKD cluster.

Management State

Cluster logging is managed by the cluster-logging-operator. It can be placed into an unmanaged state allowing an administrator to assume full control of individual component configurations and upgrades by changing the managementState from Managed to Unmanaged. Note: An unmanaged deployment will not receive updates until the ClusterLogging custom resource is placed back into a managed state.

  spec:
    managementState: "Managed"

See the unmanaged documentation for additional information.

Component Images

The image for each subcomponent are overwriteable by modifying the appropriate environment variable in the cluster-logging-operator Deployment. Each variable takes a full pull spec of the image:

  • RELATED_IMAGE_FLUENTD
  • LOGFILEMETRICEXPORTER_IMAGE
  • VECTOR_IMAGE

Common Configurations

The following configuration options apply generally to all components defined in for a ClusterLogging object (e.g. logStore, visualization, etc).

Memory and CPU

Each component specification allows for adjustments to both the CPU and memory limits. This is defined by modifying the resources block with valid memory (e.g. 16Gi) and CPU values (e.g 1):

  spec:
    logStore:
      type: "elasticsearch"
      elasticsearch:
        resources:
          limits:
            cpu:
            memory:
          requests:
            cpu:
            memory:

Node Selectors

Each component specification allows the component to target a specific node. This is defined by modifying the nodeSelector block to include key/value pairs that correspond to specifically labeled nodes:

  spec:
    logStore:
      type: "elasticsearch"
      elasticsearch:
        nodeSelector:
          node-type: infra
          cluster-logging-component: es

Data Aggregation and Storage

An Elasticsearch cluster is responsible for log aggregation. Following is a sample of the spec:

  spec:
    logStore:
      type: "elasticsearch"
      elasticsearch:
        nodeCount: 3
        storage:
          storageClassName: "gp2"
          size: "200G"
        redundancyPolicy: "SingleRedundancy"

This example specifies each data node in the cluster will be bound to a PersistentVolumeClaim that requests "200G" of "gp2" storage. Additionally, each primary shard will be backed by a single replica.

Backing storage

The cluster-logging-operator will create a PersistentVolumeClaim for each data node in the Elasticsearch cluster using the size and storage class name from the spec.

Note: Omission of the storage block will result in a deployment that includes ephemeral storage only. E.g.:

  spec:
    logStore:
      type: "elasticsearch"
      elasticsearch:
        resources
        nodeCount: 3
        redundancyPolicy: "SingleRedundancy"

Data Redundancy

The policy that defines how shards are replicated across data nodes in the cluster

Policy Description
FullRedundancy The shards for each index are fully replicated to every data node
MultipleRedundancy The shards for each index are spread over half of the data nodes
SingleRedundancy A single copy of each shard. Logs are always available and recoverable as long as at least two data nodes exist
ZeroRedundancy No copies of any shards. Logs may be unavailable (or lost) in the event a node is down or fails

Log Collectors

Log collectors are deployed as a Daemonset to each node in the OKD cluster. Following are the supported log collectors for Cluster Logging:

  • Fluentd - The default log collector based on Fluentd.
  • Vector - An alterate log collector based on Vector.
  spec:
    collection:
      type: "fluentd"
      resources:
        limits:
          cpu:
          memory:
        requests:
          cpu:
          memory:

Kibana and Visualization

Kibana is fronted by an oauth-proxy container, which additionally allows memory and CPU configuration.

  spec:
    visualization:
      type: "kibana"
      kibana:
        replicas: 1
        resources:
        proxy:
          resources:
            limits:
              cpu:
              memory:
            requests:
              cpu:
              memory:

Curator and Data Curation

Note: Curator is deprecated and no longer supported