Skip to content

Latest commit

 

History

History
109 lines (91 loc) · 5.95 KB

README.md

File metadata and controls

109 lines (91 loc) · 5.95 KB
Raw

ocp_logging role

This role allows enabling the OCP logging subsystem. This allows to aggregate cluster's metrics, application and infrastructure logs. The main components of the stack are:

  • Loki - for logs storage in an Object Storage
  • Cluster logging - for collecting for logs and metrics
  • Object storage bucket - the permanent logs storage
  • Physical volumes - local storage used for data catching

Main role tasks:

  • Validates the logging subsystem requirements
  • Creates the required secrets
  • Validates that the CustomResources are present
  • Deploys Loki stack
  • Creates a ClusterLogging instance
  • Configures the cluster to forward metrics and logs to the Loki stack
  • Enables the logging-view-plugin for logs visualization for the OCP console

This role only works on OCP 4.10 and newer and with a S3 compatible storage backend.

Variables

Variable Default Required Description
ol_access_key_id undefined Yes Key ID for the Object storage system
ol_access_key_secret undefined Yes Key Secret for the Object Storage system
ol_bucket undefined Yes Object Storage bucket name
ol_endpoint undefined Yes Object Storage endpoint
ol_region undefined Yes Object Storage region
ol_loki_size undefined Yes Loki Deployment Size. See Sizing for more details
ol_storage_class undefined Yes Cluster Storage class for Loki components
ol_event_router_image registry.redhat.io/openshift-logging/eventrouter-rhel8:v5.2.1-1 No Event Router image
ol_action install No Controls the action performed by the role: Install or cleanup
ol_namespace openshift-logging No Target namespace. It is the recommended namespace
ol_settings "" No An optional yaml file with the variables listed above. The variables defined there take precedence over the ones defined at role level

Role requirements

  • Cluster-Logging operator already installed
  • Loki-operator already installed
  • Supported Log Store (AWS S3, Google Cloud Storage, Azure, Swift, Minio, OpenShift Data Foundation) credentials (access_key and access_key_secret)
  • Logs Store endpoint address

Usage example

See below for some examples of how to use the ocp_logging role to configure the logging subsystem.

Passing the variables at role level:

- name: "Setup OCP logging stack"
  ansible.builtin.include_role:
    name: redhatci.ocp.ocp_logging
  vars:
    ol_access_key_id: <ACCESS_KEY_ID>
    ol_access_key_secret: <ACCESS_KEY_SECRET>
    ol_bucket: loki
    ol_endpoint: http://192.168.16.10:9000
    ol_region: us-east-1
    ol_loki_size: 1x.extra-small
    ol_storage_class: ocs-storagecluster-ceph-rbd
    ol_event_router_image: "registry.redhat.io/openshift-logging/eventrouter-rhel8:v5.2.1-1"

Passing the variables using a yaml vars file:

- name: "Setup OCP logging stack"
  ansible.builtin.include_role:
    name: redhatci.ocp.ocp_logging
  vars:
    ol_settings: <path_to_vars_filel.yml>

Remove resources created by the role.

- name: "Setup OCP logging stack"
  vars:
    ol_action: 'cleanup'
  ansible.builtin.include_role:
    name: redhatci.ocp.ocp_logging

Validation

To confirm that the stack is working properly:

  1. Validate that a "Logs" entry has been added to the OCP console under the "Observe" menu.
  2. Verify that the Object Storage bucket has started receiving files.

Troubleshooting

The following are some recommended actions in case there are issues during the deployment.

  1. Confirm that all the pods in the openshift-logging namespace are running.
  2. Confirm that the OCP cluster is able to reach the Object Storage bucket. Check the ingester pods for errors when flushing logs tables. 
    $ oc logs -n openshift-logging -f  logging-loki-ingester-0 | grep flush
level=info ts=2023-06-13T23:55:50.410697776Z caller=flush.go:168 msg="flushing stream" user=infrastructure fp=8532775eba54c8ba immediate=false num_chunks=1 labels="{kubernetes_container_name=\"control-plane-machine-set-operator\", kubernetes_host=\"dciokd-master-2\", kubernetes_namespace_name=\"openshift-machine-api\",
  3. Review that the Console CustomResource has patched to add enable the logging-view-plugin. 
    $ oc get console cluster -o yaml | grep spec -A4
spec:
  plugins:
    - logging-view-plugin

References