Uploaded image for project: 'OpenShift Logging'
  1. OpenShift Logging
  2. LOG-4779

Deploy collector as a deployment instead of daemonset

XMLWordPrintable

    • 5
    • False
    • None
    • False
    • NEW
    • API, Deploy
    • NEW
    • Hide
      This enhancement improves collector resource utilization by deploying collectors as a deployment with two replicas when the only input source defined in the ClusterLogForwarder is a receiver input, rather than using a daemonset on all nodes. Additionally, collectors deployed in this manner do not mount the host file system. To enable this enhancement, the ClusterLogForwarder must be annotated with `logging.openshift.io/dev-preview-enable-collector-as-deployment`.
      Show
      This enhancement improves collector resource utilization by deploying collectors as a deployment with two replicas when the only input source defined in the ClusterLogForwarder is a receiver input, rather than using a daemonset on all nodes. Additionally, collectors deployed in this manner do not mount the host file system. To enable this enhancement, the ClusterLogForwarder must be annotated with `logging.openshift.io/dev-preview-enable-collector-as-deployment`.
    • Technology Preview
    • Log Collection - Sprint 244, Log Collection - Sprint 245, Log Collection - Sprint 246, Log Collection - Sprint 247, Log Collection - Sprint 248

      Summary

      As a user of only a receiver input,
      I want to deploy the log collector as a deployment instead of daemonset
      Because I do not require the collector to run on every node

      Acceptance Criteria

      • Verify audit logs are received via the receiver and forwarded to configured outputs
      • Verify the pod no longer mounts the host file system
      • Verify the collector is a deployment when only a receiver input is referenced in the pipelines
      • Verify the collector is a daemonset when a receiver input and other built-in inputs (application, infrastructure, audit) are defined in pipelines
      • Ensure feature is behind a feature-gate

      Notes

      • This addresses the HCP usecase that utilizes one collector per NS for collecting audit logs via a webhook
      • Saves deployment resources
      • Collector does not need to mount the host file system given it only receives logs via the receiver

            rh-ee-calee Calvin Lee
            jcantril@redhat.com Jeffrey Cantrill
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated:
              Resolved: